{ "best_global_step": null, "best_metric": null, "best_model_checkpoint": null, "epoch": 0.9999146539216524, "eval_steps": 500, "global_step": 2929, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2030.0, "completions/mean_length": 1750.845703125, "completions/mean_terminated_length": 1302.2010498046875, "completions/min_length": 116.0, "completions/min_terminated_length": 116.0, "epoch": 0.0003413843133907997, "frac_reward_zero_std": 0.59375, "grad_norm": 0.05033346607156781, "learning_rate": 0.0, "loss": 0.0074, "num_tokens": 975041.0, "reward": 0.072265625, "reward_std": 0.1443791538476944, "rewards/accuracy_reward/mean": 0.072265625, "rewards/accuracy_reward/std": 0.2591804563999176, "step": 1 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.66015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2044.0, "completions/mean_length": 1776.15625, "completions/mean_terminated_length": 1248.0919189453125, "completions/min_length": 210.0, "completions/min_terminated_length": 210.0, "epoch": 0.0006827686267815994, "frac_reward_zero_std": 0.75, "grad_norm": 0.049110022117787845, "learning_rate": 3.4129692832764506e-09, "loss": 0.0006, "num_tokens": 1957665.0, "reward": 0.033203125, "reward_std": 0.0756097063422203, "rewards/accuracy_reward/mean": 0.033203125, "rewards/accuracy_reward/std": 0.17934183776378632, "step": 2 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.59765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2048.0, "completions/mean_length": 1758.369140625, "completions/mean_terminated_length": 1328.1407470703125, "completions/min_length": 215.0, "completions/min_terminated_length": 215.0, "epoch": 0.001024152940172399, "frac_reward_zero_std": 0.71875, "grad_norm": 0.04241046785362354, "learning_rate": 6.825938566552901e-09, "loss": 0.0081, "num_tokens": 2939822.0, "reward": 0.080078125, "reward_std": 0.1211543008685112, "rewards/accuracy_reward/mean": 0.080078125, "rewards/accuracy_reward/std": 0.271679550409317, "step": 3 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.69140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2034.0, "completions/mean_length": 1809.0078125, "completions/mean_terminated_length": 1273.5443115234375, "completions/min_length": 51.0, "completions/min_terminated_length": 51.0, "epoch": 0.0013655372535631989, "frac_reward_zero_std": 0.78125, "grad_norm": 0.04091706716817529, "learning_rate": 1.023890784982935e-08, "loss": 0.0069, "num_tokens": 3946434.0, "reward": 0.041015625, "reward_std": 0.07328042387962341, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 4 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.587890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2048.0, "completions/mean_length": 1707.787109375, "completions/mean_terminated_length": 1222.459716796875, "completions/min_length": 347.0, "completions/min_terminated_length": 347.0, "epoch": 0.0017069215669539984, "frac_reward_zero_std": 0.5625, "grad_norm": 0.05690206911194408, "learning_rate": 1.3651877133105802e-08, "loss": 0.022, "num_tokens": 4902325.0, "reward": 0.109375, "reward_std": 0.1755613535642624, "rewards/accuracy_reward/mean": 0.11290322244167328, "rewards/accuracy_reward/std": 0.3167939782142639, "step": 5 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1717.732421875, "completions/mean_terminated_length": 1219.3236083984375, "completions/min_length": 42.0, "completions/min_terminated_length": 42.0, "epoch": 0.002048305880344798, "frac_reward_zero_std": 0.71875, "grad_norm": 0.05573824427783701, "learning_rate": 1.706484641638225e-08, "loss": 0.0135, "num_tokens": 5869084.0, "reward": 0.041015625, "reward_std": 0.09644745290279388, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 6 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.642578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2035.0, "completions/mean_length": 1754.251953125, "completions/mean_terminated_length": 1226.1474609375, "completions/min_length": 88.0, "completions/min_terminated_length": 88.0, "epoch": 0.0023896901937355977, "frac_reward_zero_std": 0.84375, "grad_norm": 0.03002513688396799, "learning_rate": 2.04778156996587e-08, "loss": 0.0121, "num_tokens": 6839645.0, "reward": 0.044921875, "reward_std": 0.060957908630371094, "rewards/accuracy_reward/mean": 0.0463709682226181, "rewards/accuracy_reward/std": 0.21049949526786804, "step": 7 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.599609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2045.0, "completions/mean_length": 1797.53125, "completions/mean_terminated_length": 1422.43896484375, "completions/min_length": 300.0, "completions/min_terminated_length": 300.0, "epoch": 0.0027310745071263977, "frac_reward_zero_std": 0.5625, "grad_norm": 0.060326288268162755, "learning_rate": 2.3890784982935154e-08, "loss": 0.0274, "num_tokens": 7836685.0, "reward": 0.123046875, "reward_std": 0.18169808387756348, "rewards/accuracy_reward/mean": 0.123046875, "rewards/accuracy_reward/std": 0.32881227135658264, "step": 8 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.595703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1707.265625, "completions/mean_terminated_length": 1205.2174072265625, "completions/min_length": 219.0, "completions/min_terminated_length": 219.0, "epoch": 0.0030724588205171973, "frac_reward_zero_std": 0.6875, "grad_norm": 0.050131841466508904, "learning_rate": 2.7303754266211605e-08, "loss": 0.026, "num_tokens": 8784597.0, "reward": 0.048828125, "reward_std": 0.10684061795473099, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 9 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.63671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2031.0, "completions/mean_length": 1774.05859375, "completions/mean_terminated_length": 1293.9246826171875, "completions/min_length": 376.0, "completions/min_terminated_length": 376.0, "epoch": 0.003413843133907997, "frac_reward_zero_std": 0.71875, "grad_norm": 0.041968355334221295, "learning_rate": 3.071672354948805e-08, "loss": 0.0199, "num_tokens": 9764851.0, "reward": 0.0390625, "reward_std": 0.0969453901052475, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 10 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.591796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1727.42578125, "completions/mean_terminated_length": 1262.6697998046875, "completions/min_length": 241.0, "completions/min_terminated_length": 241.0, "epoch": 0.003755227447298797, "frac_reward_zero_std": 0.75, "grad_norm": 0.04477430838067097, "learning_rate": 3.41296928327645e-08, "loss": 0.0079, "num_tokens": 10724877.0, "reward": 0.05859375, "reward_std": 0.09534769505262375, "rewards/accuracy_reward/mean": 0.05859375, "rewards/accuracy_reward/std": 0.23509246110916138, "step": 11 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.59375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2037.0, "completions/mean_length": 1760.837890625, "completions/mean_terminated_length": 1341.1395263671875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.004096611760689596, "frac_reward_zero_std": 0.75, "grad_norm": 0.04142656941993241, "learning_rate": 3.754266211604096e-08, "loss": 0.0103, "num_tokens": 11697802.0, "reward": 0.07421875, "reward_std": 0.09876786172389984, "rewards/accuracy_reward/mean": 0.07421875, "rewards/accuracy_reward/std": 0.2623828947544098, "step": 12 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.64453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2023.0, "completions/mean_length": 1716.64453125, "completions/mean_terminated_length": 1115.835205078125, "completions/min_length": 127.0, "completions/min_terminated_length": 127.0, "epoch": 0.004437996074080396, "frac_reward_zero_std": 0.71875, "grad_norm": 0.043864214373648215, "learning_rate": 4.09556313993174e-08, "loss": 0.0081, "num_tokens": 12652532.0, "reward": 0.08203125, "reward_std": 0.1194583922624588, "rewards/accuracy_reward/mean": 0.08203125, "rewards/accuracy_reward/std": 0.2746807038784027, "step": 13 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.595703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.0, "completions/mean_length": 1758.34375, "completions/mean_terminated_length": 1331.5555419921875, "completions/min_length": 192.0, "completions/min_terminated_length": 192.0, "epoch": 0.0047793803874711955, "frac_reward_zero_std": 0.84375, "grad_norm": 0.03417215435914826, "learning_rate": 4.436860068259386e-08, "loss": 0.0078, "num_tokens": 13630068.0, "reward": 0.05859375, "reward_std": 0.06805649399757385, "rewards/accuracy_reward/mean": 0.05859375, "rewards/accuracy_reward/std": 0.23509246110916138, "step": 14 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.5625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2008.0, "completions/mean_length": 1646.5703125, "completions/mean_terminated_length": 1130.446533203125, "completions/min_length": 46.0, "completions/min_terminated_length": 46.0, "epoch": 0.005120764700861995, "frac_reward_zero_std": 0.71875, "grad_norm": 0.04510596346055707, "learning_rate": 4.778156996587031e-08, "loss": 0.0198, "num_tokens": 14550264.0, "reward": 0.0859375, "reward_std": 0.12361770868301392, "rewards/accuracy_reward/mean": 0.0859375, "rewards/accuracy_reward/std": 0.28054583072662354, "step": 15 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.638671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2030.0, "completions/mean_length": 1752.3125, "completions/mean_terminated_length": 1229.6649169921875, "completions/min_length": 126.0, "completions/min_terminated_length": 126.0, "epoch": 0.0054621490142527955, "frac_reward_zero_std": 0.5625, "grad_norm": 0.054358656074527854, "learning_rate": 5.119453924914675e-08, "loss": 0.023, "num_tokens": 15522616.0, "reward": 0.119140625, "reward_std": 0.16773086786270142, "rewards/accuracy_reward/mean": 0.119140625, "rewards/accuracy_reward/std": 0.32427072525024414, "step": 16 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.62890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.0, "completions/mean_length": 1725.244140625, "completions/mean_terminated_length": 1178.2579345703125, "completions/min_length": 161.0, "completions/min_terminated_length": 161.0, "epoch": 0.005803533327643595, "frac_reward_zero_std": 0.75, "grad_norm": 0.04532171535738384, "learning_rate": 5.460750853242321e-08, "loss": 0.0112, "num_tokens": 16481669.0, "reward": 0.06640625, "reward_std": 0.09589291363954544, "rewards/accuracy_reward/mean": 0.06640625, "rewards/accuracy_reward/std": 0.2492343932390213, "step": 17 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.673828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2040.0, "completions/mean_length": 1799.037109375, "completions/mean_terminated_length": 1284.712646484375, "completions/min_length": 20.0, "completions/min_terminated_length": 20.0, "epoch": 0.006144917641034395, "frac_reward_zero_std": 0.75, "grad_norm": 0.04252846636231565, "learning_rate": 5.802047781569966e-08, "loss": 0.0188, "num_tokens": 17485736.0, "reward": 0.0625, "reward_std": 0.08450747281312943, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.2422981858253479, "step": 18 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.0, "completions/mean_length": 1799.37890625, "completions/mean_terminated_length": 1356.184814453125, "completions/min_length": 360.0, "completions/min_terminated_length": 360.0, "epoch": 0.006486301954425194, "frac_reward_zero_std": 0.71875, "grad_norm": 0.044007690247254184, "learning_rate": 6.14334470989761e-08, "loss": 0.01, "num_tokens": 18484362.0, "reward": 0.068359375, "reward_std": 0.1117907464504242, "rewards/accuracy_reward/mean": 0.068359375, "rewards/accuracy_reward/std": 0.25260838866233826, "step": 19 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.58984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2043.0, "completions/mean_length": 1708.611328125, "completions/mean_terminated_length": 1220.5380859375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.006827686267815994, "frac_reward_zero_std": 0.78125, "grad_norm": 0.04488698234518163, "learning_rate": 6.484641638225255e-08, "loss": 0.0103, "num_tokens": 19436387.0, "reward": 0.08984375, "reward_std": 0.09848526120185852, "rewards/accuracy_reward/mean": 0.08984375, "rewards/accuracy_reward/std": 0.2862374484539032, "step": 20 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2041.0, "completions/mean_length": 1770.4296875, "completions/mean_terminated_length": 1307.8125, "completions/min_length": 581.0, "completions/min_terminated_length": 581.0, "epoch": 0.007169070581206793, "frac_reward_zero_std": 0.65625, "grad_norm": 0.04861825600744672, "learning_rate": 6.8259385665529e-08, "loss": 0.0229, "num_tokens": 20416207.0, "reward": 0.087890625, "reward_std": 0.13715961575508118, "rewards/accuracy_reward/mean": 0.087890625, "rewards/accuracy_reward/std": 0.2834126651287079, "step": 21 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.603515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2043.0, "completions/mean_length": 1756.564453125, "completions/mean_terminated_length": 1312.95068359375, "completions/min_length": 97.0, "completions/min_terminated_length": 97.0, "epoch": 0.007510454894597594, "frac_reward_zero_std": 0.65625, "grad_norm": 0.052662806445118816, "learning_rate": 7.167235494880546e-08, "loss": 0.0276, "num_tokens": 21398480.0, "reward": 0.107421875, "reward_std": 0.1405652016401291, "rewards/accuracy_reward/mean": 0.107421875, "rewards/accuracy_reward/std": 0.30995169281959534, "step": 22 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.662109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2040.0, "completions/mean_length": 1814.40234375, "completions/mean_terminated_length": 1356.658935546875, "completions/min_length": 95.0, "completions/min_terminated_length": 95.0, "epoch": 0.007851839207988393, "frac_reward_zero_std": 0.6875, "grad_norm": 0.04479486754650139, "learning_rate": 7.508532423208192e-08, "loss": 0.0178, "num_tokens": 22399118.0, "reward": 0.09375, "reward_std": 0.12595151364803314, "rewards/accuracy_reward/mean": 0.09375, "rewards/accuracy_reward/std": 0.29176566004753113, "step": 23 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.630859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2039.0, "completions/mean_length": 1717.69921875, "completions/mean_terminated_length": 1153.2169189453125, "completions/min_length": 171.0, "completions/min_terminated_length": 171.0, "epoch": 0.008193223521379193, "frac_reward_zero_std": 0.875, "grad_norm": 0.03288501841840136, "learning_rate": 7.849829351535836e-08, "loss": 0.0085, "num_tokens": 23358196.0, "reward": 0.04296875, "reward_std": 0.0457378551363945, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 24 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.607421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2044.0, "completions/mean_length": 1744.763671875, "completions/mean_terminated_length": 1275.5770263671875, "completions/min_length": 34.0, "completions/min_terminated_length": 34.0, "epoch": 0.008534607834769992, "frac_reward_zero_std": 0.75, "grad_norm": 0.041978908017609194, "learning_rate": 8.19112627986348e-08, "loss": 0.0196, "num_tokens": 24328619.0, "reward": 0.07421875, "reward_std": 0.09606722742319107, "rewards/accuracy_reward/mean": 0.07421875, "rewards/accuracy_reward/std": 0.2623828947544098, "step": 25 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.5390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2029.0, "completions/mean_length": 1626.05859375, "completions/mean_terminated_length": 1132.6016845703125, "completions/min_length": 218.0, "completions/min_terminated_length": 218.0, "epoch": 0.008875992148160792, "frac_reward_zero_std": 0.53125, "grad_norm": 0.062100550345779035, "learning_rate": 8.532423208191126e-08, "loss": 0.0166, "num_tokens": 25239705.0, "reward": 0.13671875, "reward_std": 0.18890517950057983, "rewards/accuracy_reward/mean": 0.1411290317773819, "rewards/accuracy_reward/std": 0.3485061228275299, "step": 26 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.5390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.0, "completions/mean_length": 1607.740234375, "completions/mean_terminated_length": 1092.8602294921875, "completions/min_length": 32.0, "completions/min_terminated_length": 32.0, "epoch": 0.009217376461551591, "frac_reward_zero_std": 0.78125, "grad_norm": 0.038504286857851476, "learning_rate": 8.873720136518772e-08, "loss": 0.0141, "num_tokens": 26132148.0, "reward": 0.021484375, "reward_std": 0.06425705552101135, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 27 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.650390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2008.0, "completions/mean_length": 1741.521484375, "completions/mean_terminated_length": 1171.36865234375, "completions/min_length": 300.0, "completions/min_terminated_length": 300.0, "epoch": 0.009558760774942391, "frac_reward_zero_std": 0.6875, "grad_norm": 0.047420119841969456, "learning_rate": 9.215017064846416e-08, "loss": 0.0179, "num_tokens": 27112895.0, "reward": 0.048828125, "reward_std": 0.11079943180084229, "rewards/accuracy_reward/mean": 0.05040322616696358, "rewards/accuracy_reward/std": 0.21899642050266266, "step": 28 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.685546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2030.0, "completions/mean_length": 1787.126953125, "completions/mean_terminated_length": 1218.391357421875, "completions/min_length": 215.0, "completions/min_terminated_length": 215.0, "epoch": 0.00990014508833319, "frac_reward_zero_std": 0.84375, "grad_norm": 0.03030950325457361, "learning_rate": 9.556313993174062e-08, "loss": 0.0093, "num_tokens": 28103168.0, "reward": 0.0546875, "reward_std": 0.056411758065223694, "rewards/accuracy_reward/mean": 0.05645161122083664, "rewards/accuracy_reward/std": 0.23102475702762604, "step": 29 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.63671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.0, "completions/mean_length": 1760.376953125, "completions/mean_terminated_length": 1256.263427734375, "completions/min_length": 280.0, "completions/min_terminated_length": 280.0, "epoch": 0.01024152940172399, "frac_reward_zero_std": 0.78125, "grad_norm": 0.03741549775073281, "learning_rate": 9.897610921501706e-08, "loss": 0.0107, "num_tokens": 29084449.0, "reward": 0.125, "reward_std": 0.08109388500452042, "rewards/accuracy_reward/mean": 0.125, "rewards/accuracy_reward/std": 0.3310423493385315, "step": 30 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.595703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2036.0, "completions/mean_length": 1744.7421875, "completions/mean_terminated_length": 1297.9130859375, "completions/min_length": 275.0, "completions/min_terminated_length": 275.0, "epoch": 0.01058291371511479, "frac_reward_zero_std": 0.59375, "grad_norm": 0.05665272966941569, "learning_rate": 1.023890784982935e-07, "loss": 0.0223, "num_tokens": 30056541.0, "reward": 0.10546875, "reward_std": 0.15262386202812195, "rewards/accuracy_reward/mean": 0.10546875, "rewards/accuracy_reward/std": 0.3074568510055542, "step": 31 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.58203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2043.0, "completions/mean_length": 1744.771484375, "completions/mean_terminated_length": 1322.5186767578125, "completions/min_length": 274.0, "completions/min_terminated_length": 274.0, "epoch": 0.010924298028505591, "frac_reward_zero_std": 0.71875, "grad_norm": 0.04584619056826458, "learning_rate": 1.0580204778156996e-07, "loss": 0.0298, "num_tokens": 31027288.0, "reward": 0.0703125, "reward_std": 0.12140804529190063, "rewards/accuracy_reward/mean": 0.0703125, "rewards/accuracy_reward/std": 0.25592297315597534, "step": 32 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.67578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1746.130859375, "completions/mean_terminated_length": 1116.9337158203125, "completions/min_length": 334.0, "completions/min_terminated_length": 334.0, "epoch": 0.01126568234189639, "frac_reward_zero_std": 0.84375, "grad_norm": 0.03438908031100313, "learning_rate": 1.0921501706484642e-07, "loss": 0.007, "num_tokens": 32006043.0, "reward": 0.0390625, "reward_std": 0.06299237906932831, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 33 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2045.0, "completions/mean_length": 1743.775390625, "completions/mean_terminated_length": 1269.1849365234375, "completions/min_length": 254.0, "completions/min_terminated_length": 254.0, "epoch": 0.01160706665528719, "frac_reward_zero_std": 0.75, "grad_norm": 0.0473121665769505, "learning_rate": 1.1262798634812286e-07, "loss": 0.0165, "num_tokens": 32972136.0, "reward": 0.056640625, "reward_std": 0.09433712065219879, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 34 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.541015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2031.0, "completions/mean_length": 1680.3125, "completions/mean_terminated_length": 1246.91064453125, "completions/min_length": 151.0, "completions/min_terminated_length": 151.0, "epoch": 0.01194845096867799, "frac_reward_zero_std": 0.6875, "grad_norm": 0.046558777905087036, "learning_rate": 1.1604095563139932e-07, "loss": 0.0117, "num_tokens": 33914472.0, "reward": 0.095703125, "reward_std": 0.12388914078474045, "rewards/accuracy_reward/mean": 0.095703125, "rewards/accuracy_reward/std": 0.2944713830947876, "step": 35 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2043.0, "completions/mean_length": 1696.216796875, "completions/mean_terminated_length": 1214.1435546875, "completions/min_length": 90.0, "completions/min_terminated_length": 90.0, "epoch": 0.01228983528206879, "frac_reward_zero_std": 0.75, "grad_norm": 0.041928760536107765, "learning_rate": 1.1945392491467578e-07, "loss": 0.0105, "num_tokens": 34866631.0, "reward": 0.041015625, "reward_std": 0.08648413419723511, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 36 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.0, "completions/mean_length": 1830.650390625, "completions/mean_terminated_length": 1385.6011962890625, "completions/min_length": 248.0, "completions/min_terminated_length": 248.0, "epoch": 0.012631219595459589, "frac_reward_zero_std": 0.78125, "grad_norm": 0.03890801617300073, "learning_rate": 1.228668941979522e-07, "loss": 0.0066, "num_tokens": 35878868.0, "reward": 0.0625, "reward_std": 0.09350419044494629, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.2422981858253479, "step": 37 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.638671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2033.0, "completions/mean_length": 1743.234375, "completions/mean_terminated_length": 1204.54052734375, "completions/min_length": 371.0, "completions/min_terminated_length": 371.0, "epoch": 0.012972603908850388, "frac_reward_zero_std": 0.75, "grad_norm": 0.03957306765731773, "learning_rate": 1.2627986348122866e-07, "loss": 0.0086, "num_tokens": 36852124.0, "reward": 0.041015625, "reward_std": 0.09110259264707565, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 38 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.642578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1774.359375, "completions/mean_terminated_length": 1282.404296875, "completions/min_length": 244.0, "completions/min_terminated_length": 244.0, "epoch": 0.013313988222241188, "frac_reward_zero_std": 0.625, "grad_norm": 0.05463734396955025, "learning_rate": 1.296928327645051e-07, "loss": 0.028, "num_tokens": 37838500.0, "reward": 0.08203125, "reward_std": 0.12384280562400818, "rewards/accuracy_reward/mean": 0.08203125, "rewards/accuracy_reward/std": 0.2746807038784027, "step": 39 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.736328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2042.0, "completions/mean_length": 1852.732421875, "completions/mean_terminated_length": 1307.4295654296875, "completions/min_length": 270.0, "completions/min_terminated_length": 270.0, "epoch": 0.013655372535631987, "frac_reward_zero_std": 0.71875, "grad_norm": 0.04512206696134928, "learning_rate": 1.3310580204778158e-07, "loss": 0.0088, "num_tokens": 38867691.0, "reward": 0.046875, "reward_std": 0.0908287987112999, "rewards/accuracy_reward/mean": 0.046875, "rewards/accuracy_reward/std": 0.21157780289649963, "step": 40 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.607421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2033.0, "completions/mean_length": 1745.134765625, "completions/mean_terminated_length": 1276.5223388671875, "completions/min_length": 426.0, "completions/min_terminated_length": 426.0, "epoch": 0.013996756849022787, "frac_reward_zero_std": 0.6875, "grad_norm": 0.04932882394737749, "learning_rate": 1.36518771331058e-07, "loss": 0.013, "num_tokens": 39832064.0, "reward": 0.037109375, "reward_std": 0.09809703379869461, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 41 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.0, "completions/mean_length": 1776.326171875, "completions/mean_terminated_length": 1199.8475341796875, "completions/min_length": 129.0, "completions/min_terminated_length": 129.0, "epoch": 0.014338141162413586, "frac_reward_zero_std": 0.65625, "grad_norm": 0.061108309601674154, "learning_rate": 1.3993174061433446e-07, "loss": 0.0093, "num_tokens": 40826903.0, "reward": 0.068359375, "reward_std": 0.12104017287492752, "rewards/accuracy_reward/mean": 0.068359375, "rewards/accuracy_reward/std": 0.25260838866233826, "step": 42 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.681640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2038.0, "completions/mean_length": 1788.171875, "completions/mean_terminated_length": 1231.852783203125, "completions/min_length": 294.0, "completions/min_terminated_length": 294.0, "epoch": 0.014679525475804386, "frac_reward_zero_std": 0.78125, "grad_norm": 0.03589300034594603, "learning_rate": 1.4334470989761092e-07, "loss": 0.0086, "num_tokens": 41816047.0, "reward": 0.052734375, "reward_std": 0.07894542813301086, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 43 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.62890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2040.0, "completions/mean_length": 1756.146484375, "completions/mean_terminated_length": 1261.5316162109375, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.015020909789195187, "frac_reward_zero_std": 0.625, "grad_norm": 0.0509461360080981, "learning_rate": 1.4675767918088735e-07, "loss": 0.0216, "num_tokens": 42785402.0, "reward": 0.099609375, "reward_std": 0.14260855317115784, "rewards/accuracy_reward/mean": 0.099609375, "rewards/accuracy_reward/std": 0.29977133870124817, "step": 44 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1785.208984375, "completions/mean_terminated_length": 1361.5255126953125, "completions/min_length": 223.0, "completions/min_terminated_length": 223.0, "epoch": 0.015362294102585987, "frac_reward_zero_std": 0.84375, "grad_norm": 0.03747732180213399, "learning_rate": 1.5017064846416383e-07, "loss": 0.0082, "num_tokens": 43781189.0, "reward": 0.04296875, "reward_std": 0.06233368441462517, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 45 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.666015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2048.0, "completions/mean_length": 1801.880859375, "completions/mean_terminated_length": 1311.0819091796875, "completions/min_length": 139.0, "completions/min_terminated_length": 139.0, "epoch": 0.015703678415976786, "frac_reward_zero_std": 0.75, "grad_norm": 0.04224816975820266, "learning_rate": 1.5358361774744026e-07, "loss": 0.0164, "num_tokens": 44778200.0, "reward": 0.0703125, "reward_std": 0.10864999890327454, "rewards/accuracy_reward/mean": 0.0703125, "rewards/accuracy_reward/std": 0.25592297315597534, "step": 46 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.630859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2024.0, "completions/mean_length": 1731.1796875, "completions/mean_terminated_length": 1189.7354736328125, "completions/min_length": 55.0, "completions/min_terminated_length": 55.0, "epoch": 0.016045062729367586, "frac_reward_zero_std": 0.84375, "grad_norm": 0.04174360995094499, "learning_rate": 1.5699658703071672e-07, "loss": 0.0133, "num_tokens": 45740404.0, "reward": 0.0546875, "reward_std": 0.05311024188995361, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 47 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.705078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2048.0, "completions/mean_length": 1836.826171875, "completions/mean_terminated_length": 1331.9669189453125, "completions/min_length": 191.0, "completions/min_terminated_length": 191.0, "epoch": 0.016386447042758386, "frac_reward_zero_std": 0.875, "grad_norm": 0.024488356216432854, "learning_rate": 1.6040955631399318e-07, "loss": 0.0053, "num_tokens": 46761563.0, "reward": 0.033203125, "reward_std": 0.04219770431518555, "rewards/accuracy_reward/mean": 0.033203125, "rewards/accuracy_reward/std": 0.17934183776378632, "step": 48 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.638671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.0, "completions/mean_length": 1769.5390625, "completions/mean_terminated_length": 1277.340576171875, "completions/min_length": 168.0, "completions/min_terminated_length": 168.0, "epoch": 0.016727831356149185, "frac_reward_zero_std": 0.8125, "grad_norm": 0.03501499334951841, "learning_rate": 1.638225255972696e-07, "loss": 0.0151, "num_tokens": 47740959.0, "reward": 0.046875, "reward_std": 0.07861834019422531, "rewards/accuracy_reward/mean": 0.046875, "rewards/accuracy_reward/std": 0.21157780289649963, "step": 49 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.55859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2048.0, "completions/mean_length": 1681.6953125, "completions/mean_terminated_length": 1218.1416015625, "completions/min_length": 26.0, "completions/min_terminated_length": 26.0, "epoch": 0.017069215669539985, "frac_reward_zero_std": 0.71875, "grad_norm": 0.04553192175857216, "learning_rate": 1.6723549488054606e-07, "loss": 0.021, "num_tokens": 48670051.0, "reward": 0.09765625, "reward_std": 0.1125703826546669, "rewards/accuracy_reward/mean": 0.09765625, "rewards/accuracy_reward/std": 0.29713961482048035, "step": 50 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.62109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.0, "completions/mean_length": 1785.296875, "completions/mean_terminated_length": 1354.680419921875, "completions/min_length": 319.0, "completions/min_terminated_length": 319.0, "epoch": 0.017410599982930784, "frac_reward_zero_std": 0.625, "grad_norm": 0.05871973844523566, "learning_rate": 1.7064846416382252e-07, "loss": 0.0233, "num_tokens": 49663355.0, "reward": 0.109375, "reward_std": 0.13375157117843628, "rewards/accuracy_reward/mean": 0.109375, "rewards/accuracy_reward/std": 0.31241437792778015, "step": 51 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.576171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2033.0, "completions/mean_length": 1647.58984375, "completions/mean_terminated_length": 1103.25341796875, "completions/min_length": 53.0, "completions/min_terminated_length": 53.0, "epoch": 0.017751984296321584, "frac_reward_zero_std": 0.8125, "grad_norm": 0.04705927425045244, "learning_rate": 1.7406143344709898e-07, "loss": 0.0102, "num_tokens": 50585241.0, "reward": 0.037109375, "reward_std": 0.069866843521595, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 52 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.630859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2031.0, "completions/mean_length": 1742.458984375, "completions/mean_terminated_length": 1220.291015625, "completions/min_length": 113.0, "completions/min_terminated_length": 113.0, "epoch": 0.018093368609712383, "frac_reward_zero_std": 0.65625, "grad_norm": 0.04988088049501189, "learning_rate": 1.7747440273037543e-07, "loss": 0.0094, "num_tokens": 51551540.0, "reward": 0.080078125, "reward_std": 0.12724943459033966, "rewards/accuracy_reward/mean": 0.080078125, "rewards/accuracy_reward/std": 0.271679550409317, "step": 53 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.53515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.0, "completions/mean_length": 1652.599609375, "completions/mean_terminated_length": 1197.390869140625, "completions/min_length": 55.0, "completions/min_terminated_length": 55.0, "epoch": 0.018434752923103183, "frac_reward_zero_std": 0.71875, "grad_norm": 0.04925080043249225, "learning_rate": 1.8088737201365186e-07, "loss": 0.0129, "num_tokens": 52480679.0, "reward": 0.0703125, "reward_std": 0.09820214658975601, "rewards/accuracy_reward/mean": 0.0703125, "rewards/accuracy_reward/std": 0.25592297315597534, "step": 54 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.662109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2048.0, "completions/mean_length": 1836.859375, "completions/mean_terminated_length": 1423.121337890625, "completions/min_length": 292.0, "completions/min_terminated_length": 292.0, "epoch": 0.018776137236493982, "frac_reward_zero_std": 0.6875, "grad_norm": 0.04356025838592198, "learning_rate": 1.8430034129692832e-07, "loss": 0.0173, "num_tokens": 53494015.0, "reward": 0.060546875, "reward_std": 0.11366777867078781, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 55 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1998.0, "completions/mean_length": 1599.03515625, "completions/mean_terminated_length": 1177.2803955078125, "completions/min_length": 279.0, "completions/min_terminated_length": 279.0, "epoch": 0.019117521549884782, "frac_reward_zero_std": 0.625, "grad_norm": 0.05553040771364216, "learning_rate": 1.8771331058020475e-07, "loss": 0.0182, "num_tokens": 54389921.0, "reward": 0.09375, "reward_std": 0.15114058554172516, "rewards/accuracy_reward/mean": 0.09375, "rewards/accuracy_reward/std": 0.29176566004753113, "step": 56 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.662109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2023.0, "completions/mean_length": 1784.33984375, "completions/mean_terminated_length": 1267.6878662109375, "completions/min_length": 249.0, "completions/min_terminated_length": 249.0, "epoch": 0.01945890586327558, "frac_reward_zero_std": 0.65625, "grad_norm": 0.05150373311041259, "learning_rate": 1.9112627986348124e-07, "loss": 0.0312, "num_tokens": 55381423.0, "reward": 0.078125, "reward_std": 0.12037593126296997, "rewards/accuracy_reward/mean": 0.078125, "rewards/accuracy_reward/std": 0.26863065361976624, "step": 57 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2044.0, "completions/mean_length": 1810.8515625, "completions/mean_terminated_length": 1342.0697021484375, "completions/min_length": 169.0, "completions/min_terminated_length": 169.0, "epoch": 0.01980029017666638, "frac_reward_zero_std": 0.75, "grad_norm": 0.04456284040734546, "learning_rate": 1.9453924914675767e-07, "loss": 0.0016, "num_tokens": 56400915.0, "reward": 0.044921875, "reward_std": 0.08793555945158005, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 58 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.67578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2048.0, "completions/mean_length": 1803.642578125, "completions/mean_terminated_length": 1294.3192138671875, "completions/min_length": 407.0, "completions/min_terminated_length": 407.0, "epoch": 0.02014167449005718, "frac_reward_zero_std": 0.71875, "grad_norm": 0.04798246912932706, "learning_rate": 1.9795221843003412e-07, "loss": 0.01, "num_tokens": 57392812.0, "reward": 0.06640625, "reward_std": 0.10369987040758133, "rewards/accuracy_reward/mean": 0.06640625, "rewards/accuracy_reward/std": 0.2492343932390213, "step": 59 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.556640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2044.0, "completions/mean_length": 1691.6328125, "completions/mean_terminated_length": 1244.21142578125, "completions/min_length": 257.0, "completions/min_terminated_length": 257.0, "epoch": 0.02048305880344798, "frac_reward_zero_std": 0.65625, "grad_norm": 0.05220746688153303, "learning_rate": 2.0136518771331058e-07, "loss": 0.0194, "num_tokens": 58342352.0, "reward": 0.0625, "reward_std": 0.12056370824575424, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.2422981858253479, "step": 60 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2043.0, "completions/mean_length": 1796.45703125, "completions/mean_terminated_length": 1299.220947265625, "completions/min_length": 341.0, "completions/min_terminated_length": 341.0, "epoch": 0.02082444311683878, "frac_reward_zero_std": 0.71875, "grad_norm": 0.04394834921332462, "learning_rate": 2.04778156996587e-07, "loss": 0.0229, "num_tokens": 59336266.0, "reward": 0.099609375, "reward_std": 0.11717641353607178, "rewards/accuracy_reward/mean": 0.099609375, "rewards/accuracy_reward/std": 0.29977133870124817, "step": 61 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.458984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1605.990234375, "completions/mean_terminated_length": 1231.0, "completions/min_length": 234.0, "completions/min_terminated_length": 234.0, "epoch": 0.02116582743022958, "frac_reward_zero_std": 0.6875, "grad_norm": 0.04553323110882962, "learning_rate": 2.0819112627986347e-07, "loss": 0.0171, "num_tokens": 60232469.0, "reward": 0.150390625, "reward_std": 0.1292927861213684, "rewards/accuracy_reward/mean": 0.150390625, "rewards/accuracy_reward/std": 0.35780346393585205, "step": 62 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.51171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2033.0, "completions/mean_length": 1584.740234375, "completions/mean_terminated_length": 1099.2440185546875, "completions/min_length": 151.0, "completions/min_terminated_length": 151.0, "epoch": 0.021507211743620382, "frac_reward_zero_std": 0.78125, "grad_norm": 0.04684939911874622, "learning_rate": 2.1160409556313992e-07, "loss": 0.0042, "num_tokens": 61126912.0, "reward": 0.060546875, "reward_std": 0.0830162912607193, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 63 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.576171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2044.0, "completions/mean_length": 1671.013671875, "completions/mean_terminated_length": 1158.520751953125, "completions/min_length": 252.0, "completions/min_terminated_length": 252.0, "epoch": 0.021848596057011182, "frac_reward_zero_std": 0.625, "grad_norm": 0.051864508251110504, "learning_rate": 2.1501706484641638e-07, "loss": 0.033, "num_tokens": 62061399.0, "reward": 0.11328125, "reward_std": 0.1546129435300827, "rewards/accuracy_reward/mean": 0.11328125, "rewards/accuracy_reward/std": 0.3172462284564972, "step": 64 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1997.0, "completions/mean_length": 1717.859375, "completions/mean_terminated_length": 1167.625, "completions/min_length": 226.0, "completions/min_terminated_length": 226.0, "epoch": 0.02218998037040198, "frac_reward_zero_std": 0.6875, "grad_norm": 0.046001761013168434, "learning_rate": 2.1843003412969284e-07, "loss": 0.0309, "num_tokens": 63021071.0, "reward": 0.09765625, "reward_std": 0.12735143303871155, "rewards/accuracy_reward/mean": 0.09765625, "rewards/accuracy_reward/std": 0.29713961482048035, "step": 65 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2039.0, "completions/mean_length": 1701.951171875, "completions/mean_terminated_length": 1227.736083984375, "completions/min_length": 255.0, "completions/min_terminated_length": 255.0, "epoch": 0.02253136468379278, "frac_reward_zero_std": 0.78125, "grad_norm": 0.03978169300789664, "learning_rate": 2.2184300341296927e-07, "loss": 0.0069, "num_tokens": 63969830.0, "reward": 0.0546875, "reward_std": 0.06799772381782532, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 66 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.705078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2038.0, "completions/mean_length": 1828.521484375, "completions/mean_terminated_length": 1303.8079833984375, "completions/min_length": 318.0, "completions/min_terminated_length": 318.0, "epoch": 0.02287274899718358, "frac_reward_zero_std": 0.75, "grad_norm": 0.0394878820084567, "learning_rate": 2.2525597269624572e-07, "loss": 0.0071, "num_tokens": 64974513.0, "reward": 0.052734375, "reward_std": 0.09382468461990356, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 67 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.673828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2048.0, "completions/mean_length": 1762.484375, "completions/mean_terminated_length": 1172.646728515625, "completions/min_length": 156.0, "completions/min_terminated_length": 156.0, "epoch": 0.02321413331057438, "frac_reward_zero_std": 0.71875, "grad_norm": 0.0487139884405135, "learning_rate": 2.2866894197952215e-07, "loss": 0.0195, "num_tokens": 65954201.0, "reward": 0.0703125, "reward_std": 0.09743183851242065, "rewards/accuracy_reward/mean": 0.0703125, "rewards/accuracy_reward/std": 0.25592297315597534, "step": 68 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.712890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2044.0, "completions/mean_length": 1868.390625, "completions/mean_terminated_length": 1422.4217529296875, "completions/min_length": 42.0, "completions/min_terminated_length": 42.0, "epoch": 0.02355551762396518, "frac_reward_zero_std": 0.875, "grad_norm": 0.026463022480313556, "learning_rate": 2.3208191126279864e-07, "loss": 0.0093, "num_tokens": 66995233.0, "reward": 0.015625, "reward_std": 0.03839729726314545, "rewards/accuracy_reward/mean": 0.016129031777381897, "rewards/accuracy_reward/std": 0.12609896063804626, "step": 69 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2035.0, "completions/mean_length": 1702.6015625, "completions/mean_terminated_length": 1285.7413330078125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.02389690193735598, "frac_reward_zero_std": 0.65625, "grad_norm": 0.052620471317906196, "learning_rate": 2.354948805460751e-07, "loss": 0.0199, "num_tokens": 67945605.0, "reward": 0.08203125, "reward_std": 0.13357868790626526, "rewards/accuracy_reward/mean": 0.08203125, "rewards/accuracy_reward/std": 0.2746807038784027, "step": 70 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.61328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2040.0, "completions/mean_length": 1719.845703125, "completions/mean_terminated_length": 1199.439453125, "completions/min_length": 63.0, "completions/min_terminated_length": 63.0, "epoch": 0.02423828625074678, "frac_reward_zero_std": 0.71875, "grad_norm": 0.05242549715798892, "learning_rate": 2.3890784982935155e-07, "loss": 0.0192, "num_tokens": 68902294.0, "reward": 0.052734375, "reward_std": 0.09946727007627487, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 71 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.638671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2036.0, "completions/mean_length": 1747.3359375, "completions/mean_terminated_length": 1215.8919677734375, "completions/min_length": 108.0, "completions/min_terminated_length": 108.0, "epoch": 0.02457967056413758, "frac_reward_zero_std": 0.78125, "grad_norm": 0.04516361120468605, "learning_rate": 2.42320819112628e-07, "loss": 0.007, "num_tokens": 69883058.0, "reward": 0.08203125, "reward_std": 0.08340215682983398, "rewards/accuracy_reward/mean": 0.08203125, "rewards/accuracy_reward/std": 0.2746807038784027, "step": 72 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2035.0, "completions/mean_length": 1658.490234375, "completions/mean_terminated_length": 1188.3922119140625, "completions/min_length": 368.0, "completions/min_terminated_length": 368.0, "epoch": 0.024921054877528378, "frac_reward_zero_std": 0.75, "grad_norm": 0.04914435813626726, "learning_rate": 2.457337883959044e-07, "loss": 0.0117, "num_tokens": 70818557.0, "reward": 0.064453125, "reward_std": 0.10293962806463242, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 73 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.55859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.0, "completions/mean_length": 1703.453125, "completions/mean_terminated_length": 1267.43359375, "completions/min_length": 332.0, "completions/min_terminated_length": 332.0, "epoch": 0.025262439190919177, "frac_reward_zero_std": 0.6875, "grad_norm": 0.04325843110868303, "learning_rate": 2.4914675767918084e-07, "loss": 0.006, "num_tokens": 71764597.0, "reward": 0.111328125, "reward_std": 0.1361807882785797, "rewards/accuracy_reward/mean": 0.111328125, "rewards/accuracy_reward/std": 0.31484565138816833, "step": 74 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.66015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1792.08203125, "completions/mean_terminated_length": 1294.9539794921875, "completions/min_length": 266.0, "completions/min_terminated_length": 266.0, "epoch": 0.025603823504309977, "frac_reward_zero_std": 0.71875, "grad_norm": 0.04052880308407635, "learning_rate": 2.525597269624573e-07, "loss": 0.0148, "num_tokens": 72762143.0, "reward": 0.05078125, "reward_std": 0.10270209610462189, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 75 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.623046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2029.0, "completions/mean_length": 1736.513671875, "completions/mean_terminated_length": 1221.673583984375, "completions/min_length": 312.0, "completions/min_terminated_length": 312.0, "epoch": 0.025945207817700777, "frac_reward_zero_std": 0.6875, "grad_norm": 0.0442077917273844, "learning_rate": 2.5597269624573375e-07, "loss": 0.0171, "num_tokens": 73720918.0, "reward": 0.083984375, "reward_std": 0.125734344124794, "rewards/accuracy_reward/mean": 0.083984375, "rewards/accuracy_reward/std": 0.2776356339454651, "step": 76 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.708984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2030.0, "completions/mean_length": 1820.267578125, "completions/mean_terminated_length": 1265.4564208984375, "completions/min_length": 26.0, "completions/min_terminated_length": 26.0, "epoch": 0.026286592131091576, "frac_reward_zero_std": 0.78125, "grad_norm": 0.0413919985499403, "learning_rate": 2.593856655290102e-07, "loss": 0.0086, "num_tokens": 74730399.0, "reward": 0.056640625, "reward_std": 0.08444078266620636, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 77 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.52734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2033.0, "completions/mean_length": 1644.21875, "completions/mean_terminated_length": 1193.888427734375, "completions/min_length": 18.0, "completions/min_terminated_length": 18.0, "epoch": 0.026627976444482376, "frac_reward_zero_std": 0.6875, "grad_norm": 0.05256560820797869, "learning_rate": 2.627986348122867e-07, "loss": 0.0193, "num_tokens": 75653551.0, "reward": 0.134765625, "reward_std": 0.12078087776899338, "rewards/accuracy_reward/mean": 0.1391129046678543, "rewards/accuracy_reward/std": 0.3464137017726898, "step": 78 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1755.6875, "completions/mean_terminated_length": 1268.5, "completions/min_length": 395.0, "completions/min_terminated_length": 395.0, "epoch": 0.026969360757873175, "frac_reward_zero_std": 0.8125, "grad_norm": 0.049319138504193165, "learning_rate": 2.6621160409556315e-07, "loss": 0.0061, "num_tokens": 76633087.0, "reward": 0.0234375, "reward_std": 0.062167368829250336, "rewards/accuracy_reward/mean": 0.0234375, "rewards/accuracy_reward/std": 0.15143637359142303, "step": 79 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2037.0, "completions/mean_length": 1687.9921875, "completions/mean_terminated_length": 1144.691162109375, "completions/min_length": 211.0, "completions/min_terminated_length": 211.0, "epoch": 0.027310745071263975, "frac_reward_zero_std": 0.75, "grad_norm": 0.04549203694995152, "learning_rate": 2.696245733788396e-07, "loss": 0.0162, "num_tokens": 77581675.0, "reward": 0.041015625, "reward_std": 0.0879673957824707, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 80 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.498046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2041.0, "completions/mean_length": 1614.78125, "completions/mean_terminated_length": 1184.933837890625, "completions/min_length": 294.0, "completions/min_terminated_length": 294.0, "epoch": 0.027652129384654774, "frac_reward_zero_std": 0.6875, "grad_norm": 0.04830342329842642, "learning_rate": 2.73037542662116e-07, "loss": 0.0071, "num_tokens": 78480283.0, "reward": 0.11328125, "reward_std": 0.12803566455841064, "rewards/accuracy_reward/mean": 0.11328125, "rewards/accuracy_reward/std": 0.3172462284564972, "step": 81 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2045.0, "completions/mean_length": 1852.708984375, "completions/mean_terminated_length": 1372.398681640625, "completions/min_length": 88.0, "completions/min_terminated_length": 88.0, "epoch": 0.027993513698045574, "frac_reward_zero_std": 0.625, "grad_norm": 0.05194084815029283, "learning_rate": 2.764505119453925e-07, "loss": 0.0145, "num_tokens": 79508278.0, "reward": 0.08203125, "reward_std": 0.1338978409767151, "rewards/accuracy_reward/mean": 0.08203125, "rewards/accuracy_reward/std": 0.2746807038784027, "step": 82 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.67578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2048.0, "completions/mean_length": 1789.0, "completions/mean_terminated_length": 1249.1566162109375, "completions/min_length": 15.0, "completions/min_terminated_length": 15.0, "epoch": 0.028334898011436373, "frac_reward_zero_std": 0.8125, "grad_norm": 0.04420763351849259, "learning_rate": 2.798634812286689e-07, "loss": 0.0102, "num_tokens": 80498390.0, "reward": 0.037109375, "reward_std": 0.07174387574195862, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 83 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.0, "completions/mean_length": 1737.111328125, "completions/mean_terminated_length": 1235.882568359375, "completions/min_length": 164.0, "completions/min_terminated_length": 164.0, "epoch": 0.028676282324827173, "frac_reward_zero_std": 0.6875, "grad_norm": 0.049489819333455864, "learning_rate": 2.8327645051194536e-07, "loss": 0.0108, "num_tokens": 81462431.0, "reward": 0.05859375, "reward_std": 0.1049705445766449, "rewards/accuracy_reward/mean": 0.05859375, "rewards/accuracy_reward/std": 0.23509246110916138, "step": 84 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.658203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1795.79296875, "completions/mean_terminated_length": 1310.1142578125, "completions/min_length": 215.0, "completions/min_terminated_length": 215.0, "epoch": 0.029017666638217973, "frac_reward_zero_std": 0.78125, "grad_norm": 0.03697360030207983, "learning_rate": 2.8668941979522184e-07, "loss": 0.0059, "num_tokens": 82457717.0, "reward": 0.0625, "reward_std": 0.0752037912607193, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.2422981858253479, "step": 85 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.654296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2028.0, "completions/mean_length": 1781.0, "completions/mean_terminated_length": 1275.6610107421875, "completions/min_length": 256.0, "completions/min_terminated_length": 256.0, "epoch": 0.029359050951608772, "frac_reward_zero_std": 0.78125, "grad_norm": 0.04519682757594774, "learning_rate": 2.9010238907849827e-07, "loss": 0.0183, "num_tokens": 83454821.0, "reward": 0.072265625, "reward_std": 0.07834454625844955, "rewards/accuracy_reward/mean": 0.072265625, "rewards/accuracy_reward/std": 0.2591804563999176, "step": 86 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.5546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2035.0, "completions/mean_length": 1573.21875, "completions/mean_terminated_length": 981.99560546875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.02970043526499957, "frac_reward_zero_std": 0.78125, "grad_norm": 0.0437480076324369, "learning_rate": 2.935153583617747e-07, "loss": 0.0144, "num_tokens": 84339077.0, "reward": 0.064453125, "reward_std": 0.094484344124794, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 87 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.646484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2043.0, "completions/mean_length": 1779.380859375, "completions/mean_terminated_length": 1288.3978271484375, "completions/min_length": 170.0, "completions/min_terminated_length": 170.0, "epoch": 0.030041819578390375, "frac_reward_zero_std": 0.625, "grad_norm": 0.05401587306569221, "learning_rate": 2.969283276450512e-07, "loss": 0.0219, "num_tokens": 85337224.0, "reward": 0.109375, "reward_std": 0.15756931900978088, "rewards/accuracy_reward/mean": 0.109375, "rewards/accuracy_reward/std": 0.31241437792778015, "step": 88 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.53125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2036.0, "completions/mean_length": 1696.1953125, "completions/mean_terminated_length": 1297.4833984375, "completions/min_length": 369.0, "completions/min_terminated_length": 369.0, "epoch": 0.030383203891781174, "frac_reward_zero_std": 0.78125, "grad_norm": 0.03941846800047262, "learning_rate": 3.0034129692832767e-07, "loss": 0.0144, "num_tokens": 86279324.0, "reward": 0.04296875, "reward_std": 0.0828578919172287, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 89 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.65234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2017.0, "completions/mean_length": 1741.275390625, "completions/mean_terminated_length": 1165.7359619140625, "completions/min_length": 59.0, "completions/min_terminated_length": 59.0, "epoch": 0.030724588205171974, "frac_reward_zero_std": 0.75, "grad_norm": 0.049521798619494864, "learning_rate": 3.037542662116041e-07, "loss": -0.0008, "num_tokens": 87245913.0, "reward": 0.068359375, "reward_std": 0.09969472885131836, "rewards/accuracy_reward/mean": 0.068359375, "rewards/accuracy_reward/std": 0.25260838866233826, "step": 90 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.56640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1682.029296875, "completions/mean_terminated_length": 1203.95947265625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.031065972518562773, "frac_reward_zero_std": 0.75, "grad_norm": 0.04393887888990321, "learning_rate": 3.0716723549488053e-07, "loss": 0.0182, "num_tokens": 88182408.0, "reward": 0.044921875, "reward_std": 0.09628114849328995, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 91 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.798828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2048.0, "completions/mean_length": 1910.404296875, "completions/mean_terminated_length": 1364.59228515625, "completions/min_length": 398.0, "completions/min_terminated_length": 398.0, "epoch": 0.03140735683195357, "frac_reward_zero_std": 0.8125, "grad_norm": 0.03713949427220516, "learning_rate": 3.10580204778157e-07, "loss": 0.0095, "num_tokens": 89245159.0, "reward": 0.037109375, "reward_std": 0.07117830216884613, "rewards/accuracy_reward/mean": 0.03958333283662796, "rewards/accuracy_reward/std": 0.19518160820007324, "step": 92 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.5078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2033.0, "completions/mean_length": 1593.330078125, "completions/mean_terminated_length": 1124.2261962890625, "completions/min_length": 28.0, "completions/min_terminated_length": 28.0, "epoch": 0.03174874114534437, "frac_reward_zero_std": 0.65625, "grad_norm": 0.057589385873441194, "learning_rate": 3.1399317406143344e-07, "loss": 0.0408, "num_tokens": 90130080.0, "reward": 0.087890625, "reward_std": 0.13479050993919373, "rewards/accuracy_reward/mean": 0.087890625, "rewards/accuracy_reward/std": 0.2834126651287079, "step": 93 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.70703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2044.0, "completions/mean_length": 1829.37109375, "completions/mean_terminated_length": 1301.7467041015625, "completions/min_length": 274.0, "completions/min_terminated_length": 274.0, "epoch": 0.03209012545873517, "frac_reward_zero_std": 0.84375, "grad_norm": 0.03981280813413336, "learning_rate": 3.1740614334470987e-07, "loss": 0.015, "num_tokens": 91144382.0, "reward": 0.029296875, "reward_std": 0.058760739862918854, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 94 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.638671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2048.0, "completions/mean_length": 1764.8671875, "completions/mean_terminated_length": 1264.6378173828125, "completions/min_length": 81.0, "completions/min_terminated_length": 81.0, "epoch": 0.03243150977212597, "frac_reward_zero_std": 0.65625, "grad_norm": 0.047823694560574635, "learning_rate": 3.2081911262798635e-07, "loss": 0.0279, "num_tokens": 92129738.0, "reward": 0.076171875, "reward_std": 0.1329134702682495, "rewards/accuracy_reward/mean": 0.076171875, "rewards/accuracy_reward/std": 0.26553234457969666, "step": 95 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2033.0, "completions/mean_length": 1760.23828125, "completions/mean_terminated_length": 1191.406982421875, "completions/min_length": 341.0, "completions/min_terminated_length": 341.0, "epoch": 0.03277289408551677, "frac_reward_zero_std": 0.59375, "grad_norm": 0.053190577942516325, "learning_rate": 3.242320819112628e-07, "loss": 0.0268, "num_tokens": 93100820.0, "reward": 0.078125, "reward_std": 0.1441391110420227, "rewards/accuracy_reward/mean": 0.078125, "rewards/accuracy_reward/std": 0.26863065361976624, "step": 96 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.595703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2043.0, "completions/mean_length": 1761.7578125, "completions/mean_terminated_length": 1340.0, "completions/min_length": 295.0, "completions/min_terminated_length": 295.0, "epoch": 0.03311427839890757, "frac_reward_zero_std": 0.6875, "grad_norm": 0.04518469814679981, "learning_rate": 3.276450511945392e-07, "loss": 0.0142, "num_tokens": 94078888.0, "reward": 0.087890625, "reward_std": 0.11125211417675018, "rewards/accuracy_reward/mean": 0.087890625, "rewards/accuracy_reward/std": 0.2834126651287079, "step": 97 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.677734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2000.0, "completions/mean_length": 1793.564453125, "completions/mean_terminated_length": 1258.478759765625, "completions/min_length": 334.0, "completions/min_terminated_length": 334.0, "epoch": 0.03345566271229837, "frac_reward_zero_std": 0.75, "grad_norm": 0.041203920562389446, "learning_rate": 3.3105802047781565e-07, "loss": 0.0169, "num_tokens": 95078089.0, "reward": 0.05078125, "reward_std": 0.08709507435560226, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 98 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2026.0, "completions/mean_length": 1794.5546875, "completions/mean_terminated_length": 1256.756103515625, "completions/min_length": 141.0, "completions/min_terminated_length": 141.0, "epoch": 0.03379704702568917, "frac_reward_zero_std": 0.71875, "grad_norm": 0.04788091387888716, "learning_rate": 3.3447098976109213e-07, "loss": 0.0212, "num_tokens": 96073781.0, "reward": 0.03515625, "reward_std": 0.0908288061618805, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 99 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.642578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2037.0, "completions/mean_length": 1782.1171875, "completions/mean_terminated_length": 1304.1092529296875, "completions/min_length": 184.0, "completions/min_terminated_length": 184.0, "epoch": 0.03413843133907997, "frac_reward_zero_std": 0.78125, "grad_norm": 0.04097895400370385, "learning_rate": 3.3788395904436856e-07, "loss": 0.0144, "num_tokens": 97063361.0, "reward": 0.060546875, "reward_std": 0.08279120922088623, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 100 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.72265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2043.0, "completions/mean_length": 1863.099609375, "completions/mean_terminated_length": 1381.31689453125, "completions/min_length": 331.0, "completions/min_terminated_length": 331.0, "epoch": 0.03447981565247077, "frac_reward_zero_std": 0.75, "grad_norm": 0.04778362901424674, "learning_rate": 3.4129692832764504e-07, "loss": 0.014, "num_tokens": 98088532.0, "reward": 0.05078125, "reward_std": 0.07300759106874466, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 101 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.73828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2036.0, "completions/mean_length": 1841.681640625, "completions/mean_terminated_length": 1259.6790771484375, "completions/min_length": 321.0, "completions/min_terminated_length": 321.0, "epoch": 0.03482119996586157, "frac_reward_zero_std": 0.84375, "grad_norm": 0.03397151841054294, "learning_rate": 3.447098976109215e-07, "loss": 0.0127, "num_tokens": 99108657.0, "reward": 0.025390625, "reward_std": 0.05335709825158119, "rewards/accuracy_reward/mean": 0.025390625, "rewards/accuracy_reward/std": 0.15746226906776428, "step": 102 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.615234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1732.984375, "completions/mean_terminated_length": 1229.2791748046875, "completions/min_length": 102.0, "completions/min_terminated_length": 102.0, "epoch": 0.03516258427925237, "frac_reward_zero_std": 0.78125, "grad_norm": 0.05057841585195066, "learning_rate": 3.4812286689419796e-07, "loss": 0.0141, "num_tokens": 100073801.0, "reward": 0.056640625, "reward_std": 0.08042868226766586, "rewards/accuracy_reward/mean": 0.058467742055654526, "rewards/accuracy_reward/std": 0.23486268520355225, "step": 103 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.544921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2025.0, "completions/mean_length": 1723.919921875, "completions/mean_terminated_length": 1336.0943603515625, "completions/min_length": 150.0, "completions/min_terminated_length": 150.0, "epoch": 0.03550396859264317, "frac_reward_zero_std": 0.5, "grad_norm": 0.06190229552777436, "learning_rate": 3.515358361774744e-07, "loss": 0.0281, "num_tokens": 101037888.0, "reward": 0.1484375, "reward_std": 0.21379581093788147, "rewards/accuracy_reward/mean": 0.15322580933570862, "rewards/accuracy_reward/std": 0.36056873202323914, "step": 104 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2029.0, "completions/mean_length": 1829.623046875, "completions/mean_terminated_length": 1271.548583984375, "completions/min_length": 219.0, "completions/min_terminated_length": 219.0, "epoch": 0.03584535290603397, "frac_reward_zero_std": 0.90625, "grad_norm": 0.02390952553732917, "learning_rate": 3.5494880546075087e-07, "loss": 0.0049, "num_tokens": 102051039.0, "reward": 0.037109375, "reward_std": 0.039834219962358475, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 105 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.60546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2030.0, "completions/mean_length": 1740.14453125, "completions/mean_terminated_length": 1267.876220703125, "completions/min_length": 92.0, "completions/min_terminated_length": 92.0, "epoch": 0.03618673721942477, "frac_reward_zero_std": 0.59375, "grad_norm": 0.054417121947166296, "learning_rate": 3.583617747440273e-07, "loss": 0.0105, "num_tokens": 103025513.0, "reward": 0.126953125, "reward_std": 0.16408398747444153, "rewards/accuracy_reward/mean": 0.126953125, "rewards/accuracy_reward/std": 0.33324605226516724, "step": 106 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.5, "completions/max_length": 2048.0, "completions/max_terminated_length": 2032.0, "completions/mean_length": 1633.775390625, "completions/mean_terminated_length": 1219.55078125, "completions/min_length": 120.0, "completions/min_terminated_length": 120.0, "epoch": 0.036528121532815566, "frac_reward_zero_std": 0.75, "grad_norm": 0.046098620094480694, "learning_rate": 3.6177474402730373e-07, "loss": 0.0159, "num_tokens": 103939222.0, "reward": 0.068359375, "reward_std": 0.09759023785591125, "rewards/accuracy_reward/mean": 0.068359375, "rewards/accuracy_reward/std": 0.25260838866233826, "step": 107 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2045.0, "completions/mean_length": 1798.673828125, "completions/mean_terminated_length": 1305.8197021484375, "completions/min_length": 71.0, "completions/min_terminated_length": 71.0, "epoch": 0.036869505846206366, "frac_reward_zero_std": 0.8125, "grad_norm": 0.034770213970435726, "learning_rate": 3.6518771331058016e-07, "loss": 0.015, "num_tokens": 104942143.0, "reward": 0.05859375, "reward_std": 0.08103963732719421, "rewards/accuracy_reward/mean": 0.05859375, "rewards/accuracy_reward/std": 0.23509246110916138, "step": 108 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.576171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2039.0, "completions/mean_length": 1714.1328125, "completions/mean_terminated_length": 1260.258056640625, "completions/min_length": 215.0, "completions/min_terminated_length": 215.0, "epoch": 0.037210890159597165, "frac_reward_zero_std": 0.71875, "grad_norm": 0.04945430879362447, "learning_rate": 3.6860068259385664e-07, "loss": 0.0106, "num_tokens": 105891395.0, "reward": 0.119140625, "reward_std": 0.11643067002296448, "rewards/accuracy_reward/mean": 0.119140625, "rewards/accuracy_reward/std": 0.32427072525024414, "step": 109 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.5390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1700.833984375, "completions/mean_terminated_length": 1294.8262939453125, "completions/min_length": 332.0, "completions/min_terminated_length": 332.0, "epoch": 0.037552274472987965, "frac_reward_zero_std": 0.6875, "grad_norm": 0.0476936256201125, "learning_rate": 3.7201365187713307e-07, "loss": 0.015, "num_tokens": 106837294.0, "reward": 0.080078125, "reward_std": 0.12584641575813293, "rewards/accuracy_reward/mean": 0.080078125, "rewards/accuracy_reward/std": 0.271679550409317, "step": 110 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.59375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2043.0, "completions/mean_length": 1715.486328125, "completions/mean_terminated_length": 1229.5048828125, "completions/min_length": 235.0, "completions/min_terminated_length": 235.0, "epoch": 0.037893658786378764, "frac_reward_zero_std": 0.6875, "grad_norm": 0.04945403629392554, "learning_rate": 3.754266211604095e-07, "loss": 0.0163, "num_tokens": 107792503.0, "reward": 0.052734375, "reward_std": 0.10982763022184372, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 111 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.580078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2030.0, "completions/mean_length": 1707.126953125, "completions/mean_terminated_length": 1236.2464599609375, "completions/min_length": 355.0, "completions/min_terminated_length": 355.0, "epoch": 0.038235043099769564, "frac_reward_zero_std": 0.78125, "grad_norm": 0.044412513761630794, "learning_rate": 3.78839590443686e-07, "loss": 0.026, "num_tokens": 108741480.0, "reward": 0.076171875, "reward_std": 0.09938118606805801, "rewards/accuracy_reward/mean": 0.076171875, "rewards/accuracy_reward/std": 0.26553234457969666, "step": 112 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.580078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2019.0, "completions/mean_length": 1656.69140625, "completions/mean_terminated_length": 1116.1395263671875, "completions/min_length": 250.0, "completions/min_terminated_length": 250.0, "epoch": 0.038576427413160363, "frac_reward_zero_std": 0.78125, "grad_norm": 0.04192073828764485, "learning_rate": 3.8225255972696247e-07, "loss": 0.0162, "num_tokens": 109663930.0, "reward": 0.076171875, "reward_std": 0.08862704038619995, "rewards/accuracy_reward/mean": 0.0786290317773819, "rewards/accuracy_reward/std": 0.26943066716194153, "step": 113 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2011.0, "completions/mean_length": 1618.369140625, "completions/mean_terminated_length": 1214.776611328125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.03891781172655116, "frac_reward_zero_std": 0.71875, "grad_norm": 0.052787331840070895, "learning_rate": 3.856655290102389e-07, "loss": 0.0181, "num_tokens": 110571415.0, "reward": 0.09765625, "reward_std": 0.11698848009109497, "rewards/accuracy_reward/mean": 0.09765625, "rewards/accuracy_reward/std": 0.29713961482048035, "step": 114 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.603515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1763.501953125, "completions/mean_terminated_length": 1330.4482421875, "completions/min_length": 448.0, "completions/min_terminated_length": 448.0, "epoch": 0.03925919603994196, "frac_reward_zero_std": 0.625, "grad_norm": 0.05457150666025581, "learning_rate": 3.8907849829351533e-07, "loss": 0.0262, "num_tokens": 111542552.0, "reward": 0.109375, "reward_std": 0.14895127713680267, "rewards/accuracy_reward/mean": 0.109375, "rewards/accuracy_reward/std": 0.31241437792778015, "step": 115 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.59375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2034.0, "completions/mean_length": 1690.7265625, "completions/mean_terminated_length": 1168.5577392578125, "completions/min_length": 30.0, "completions/min_terminated_length": 30.0, "epoch": 0.03960058035333276, "frac_reward_zero_std": 0.75, "grad_norm": 0.044330639053892756, "learning_rate": 3.924914675767918e-07, "loss": 0.008, "num_tokens": 112478396.0, "reward": 0.060546875, "reward_std": 0.09655951708555222, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 116 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.53125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2048.0, "completions/mean_length": 1607.65234375, "completions/mean_terminated_length": 1108.5916748046875, "completions/min_length": 61.0, "completions/min_terminated_length": 61.0, "epoch": 0.03994196466672356, "frac_reward_zero_std": 0.65625, "grad_norm": 0.05546016146959169, "learning_rate": 3.9590443686006824e-07, "loss": 0.0166, "num_tokens": 113374778.0, "reward": 0.064453125, "reward_std": 0.12579216063022614, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 117 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.501953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2045.0, "completions/mean_length": 1647.279296875, "completions/mean_terminated_length": 1243.415771484375, "completions/min_length": 392.0, "completions/min_terminated_length": 392.0, "epoch": 0.04028334898011436, "frac_reward_zero_std": 0.71875, "grad_norm": 0.04647878295869337, "learning_rate": 3.993174061433447e-07, "loss": 0.0206, "num_tokens": 114297641.0, "reward": 0.0625, "reward_std": 0.10300742089748383, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.2422981858253479, "step": 118 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.720703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2032.0, "completions/mean_length": 1825.35546875, "completions/mean_terminated_length": 1250.839111328125, "completions/min_length": 152.0, "completions/min_terminated_length": 152.0, "epoch": 0.04062473329350516, "frac_reward_zero_std": 0.71875, "grad_norm": 0.05134646675164814, "learning_rate": 4.0273037542662116e-07, "loss": 0.017, "num_tokens": 115317759.0, "reward": 0.056640625, "reward_std": 0.106694795191288, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 119 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.560546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1710.044921875, "completions/mean_terminated_length": 1278.9644775390625, "completions/min_length": 30.0, "completions/min_terminated_length": 30.0, "epoch": 0.04096611760689596, "frac_reward_zero_std": 0.65625, "grad_norm": 0.05048273340287397, "learning_rate": 4.061433447098976e-07, "loss": 0.0092, "num_tokens": 116264022.0, "reward": 0.123046875, "reward_std": 0.15681803226470947, "rewards/accuracy_reward/mean": 0.123046875, "rewards/accuracy_reward/std": 0.32881227135658264, "step": 120 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.58203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2045.0, "completions/mean_length": 1736.603515625, "completions/mean_terminated_length": 1302.9765625, "completions/min_length": 268.0, "completions/min_terminated_length": 268.0, "epoch": 0.04130750192028676, "frac_reward_zero_std": 0.71875, "grad_norm": 0.05672386558413753, "learning_rate": 4.09556313993174e-07, "loss": 0.0276, "num_tokens": 117230011.0, "reward": 0.0859375, "reward_std": 0.10480518639087677, "rewards/accuracy_reward/mean": 0.08870967477560043, "rewards/accuracy_reward/std": 0.284611314535141, "step": 121 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.630859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2043.0, "completions/mean_length": 1741.84765625, "completions/mean_terminated_length": 1218.851806640625, "completions/min_length": 8.0, "completions/min_terminated_length": 8.0, "epoch": 0.04164888623367756, "frac_reward_zero_std": 0.8125, "grad_norm": 0.038448628495947025, "learning_rate": 4.1296928327645045e-07, "loss": 0.0152, "num_tokens": 118195469.0, "reward": 0.076171875, "reward_std": 0.07191018760204315, "rewards/accuracy_reward/mean": 0.076171875, "rewards/accuracy_reward/std": 0.26553234457969666, "step": 122 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.548828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2014.0, "completions/mean_length": 1681.34375, "completions/mean_terminated_length": 1235.32470703125, "completions/min_length": 92.0, "completions/min_terminated_length": 92.0, "epoch": 0.04199027054706836, "frac_reward_zero_std": 0.625, "grad_norm": 0.054195296284921715, "learning_rate": 4.1638225255972693e-07, "loss": 0.0155, "num_tokens": 119134173.0, "reward": 0.103515625, "reward_std": 0.15227124094963074, "rewards/accuracy_reward/mean": 0.103515625, "rewards/accuracy_reward/std": 0.30492907762527466, "step": 123 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2038.0, "completions/mean_length": 1593.080078125, "completions/mean_terminated_length": 1165.7310791015625, "completions/min_length": 253.0, "completions/min_terminated_length": 253.0, "epoch": 0.04233165486045916, "frac_reward_zero_std": 0.59375, "grad_norm": 0.0607902927918571, "learning_rate": 4.1979522184300336e-07, "loss": 0.0303, "num_tokens": 120022598.0, "reward": 0.154296875, "reward_std": 0.14884379506111145, "rewards/accuracy_reward/mean": 0.154296875, "rewards/accuracy_reward/std": 0.36158639192581177, "step": 124 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2017.0, "completions/mean_length": 1650.2265625, "completions/mean_terminated_length": 1170.1551513671875, "completions/min_length": 153.0, "completions/min_terminated_length": 153.0, "epoch": 0.04267303917384996, "frac_reward_zero_std": 0.75, "grad_norm": 0.052516768559291355, "learning_rate": 4.2320819112627985e-07, "loss": 0.0168, "num_tokens": 120954314.0, "reward": 0.05078125, "reward_std": 0.09338457882404327, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 125 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.60546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2027.0, "completions/mean_length": 1744.26953125, "completions/mean_terminated_length": 1278.1484375, "completions/min_length": 163.0, "completions/min_terminated_length": 163.0, "epoch": 0.043014423487240765, "frac_reward_zero_std": 0.71875, "grad_norm": 0.05674187549515702, "learning_rate": 4.2662116040955633e-07, "loss": 0.0239, "num_tokens": 121927956.0, "reward": 0.072265625, "reward_std": 0.09126891195774078, "rewards/accuracy_reward/mean": 0.072265625, "rewards/accuracy_reward/std": 0.2591804563999176, "step": 126 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2034.0, "completions/mean_length": 1589.90625, "completions/mean_terminated_length": 1159.5758056640625, "completions/min_length": 313.0, "completions/min_terminated_length": 313.0, "epoch": 0.043355807800631564, "frac_reward_zero_std": 0.75, "grad_norm": 0.04441013244691082, "learning_rate": 4.3003412969283276e-07, "loss": 0.0253, "num_tokens": 122812660.0, "reward": 0.0859375, "reward_std": 0.1043786108493805, "rewards/accuracy_reward/mean": 0.0859375, "rewards/accuracy_reward/std": 0.28054583072662354, "step": 127 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.65625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2043.0, "completions/mean_length": 1740.98828125, "completions/mean_terminated_length": 1154.875, "completions/min_length": 196.0, "completions/min_terminated_length": 196.0, "epoch": 0.043697192114022364, "frac_reward_zero_std": 0.78125, "grad_norm": 0.03913875714149468, "learning_rate": 4.334470989761092e-07, "loss": 0.0151, "num_tokens": 123785246.0, "reward": 0.05859375, "reward_std": 0.07630911469459534, "rewards/accuracy_reward/mean": 0.05859375, "rewards/accuracy_reward/std": 0.23509246110916138, "step": 128 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.525390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2045.0, "completions/mean_length": 1636.26171875, "completions/mean_terminated_length": 1180.4691162109375, "completions/min_length": 74.0, "completions/min_terminated_length": 74.0, "epoch": 0.04403857642741316, "frac_reward_zero_std": 0.6875, "grad_norm": 0.05295970073418793, "learning_rate": 4.3686006825938567e-07, "loss": 0.0184, "num_tokens": 124703780.0, "reward": 0.134765625, "reward_std": 0.13888943195343018, "rewards/accuracy_reward/mean": 0.134765625, "rewards/accuracy_reward/std": 0.3418070077896118, "step": 129 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.486328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1618.779296875, "completions/mean_terminated_length": 1212.4068603515625, "completions/min_length": 264.0, "completions/min_terminated_length": 264.0, "epoch": 0.04437996074080396, "frac_reward_zero_std": 0.59375, "grad_norm": 0.05296995210886687, "learning_rate": 4.402730375426621e-07, "loss": 0.0167, "num_tokens": 125612419.0, "reward": 0.08984375, "reward_std": 0.15933895111083984, "rewards/accuracy_reward/mean": 0.08984375, "rewards/accuracy_reward/std": 0.2862374484539032, "step": 130 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.603515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2030.0, "completions/mean_length": 1714.74609375, "completions/mean_terminated_length": 1207.477783203125, "completions/min_length": 46.0, "completions/min_terminated_length": 46.0, "epoch": 0.04472134505419476, "frac_reward_zero_std": 0.65625, "grad_norm": 0.05022631992454979, "learning_rate": 4.4368600682593853e-07, "loss": 0.0216, "num_tokens": 126569713.0, "reward": 0.072265625, "reward_std": 0.12926754355430603, "rewards/accuracy_reward/mean": 0.072265625, "rewards/accuracy_reward/std": 0.2591804563999176, "step": 131 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.666015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2037.0, "completions/mean_length": 1803.328125, "completions/mean_terminated_length": 1315.5906982421875, "completions/min_length": 402.0, "completions/min_terminated_length": 402.0, "epoch": 0.04506272936758556, "frac_reward_zero_std": 0.6875, "grad_norm": 0.053551969848103516, "learning_rate": 4.4709897610921496e-07, "loss": 0.0277, "num_tokens": 127580665.0, "reward": 0.119140625, "reward_std": 0.12450800091028214, "rewards/accuracy_reward/mean": 0.119140625, "rewards/accuracy_reward/std": 0.32427072525024414, "step": 132 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2043.0, "completions/mean_length": 1713.65234375, "completions/mean_terminated_length": 1156.40625, "completions/min_length": 10.0, "completions/min_terminated_length": 10.0, "epoch": 0.04540411368097636, "frac_reward_zero_std": 0.78125, "grad_norm": 0.042762161364303275, "learning_rate": 4.5051194539249145e-07, "loss": 0.0082, "num_tokens": 128532471.0, "reward": 0.05078125, "reward_std": 0.0767945945262909, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 133 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.533203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2032.0, "completions/mean_length": 1646.6328125, "completions/mean_terminated_length": 1188.1673583984375, "completions/min_length": 197.0, "completions/min_terminated_length": 197.0, "epoch": 0.04574549799436716, "frac_reward_zero_std": 0.75, "grad_norm": 0.048225369539176284, "learning_rate": 4.539249146757679e-07, "loss": 0.0173, "num_tokens": 129449547.0, "reward": 0.07421875, "reward_std": 0.09341736882925034, "rewards/accuracy_reward/mean": 0.07421875, "rewards/accuracy_reward/std": 0.2623828947544098, "step": 134 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2048.0, "completions/mean_length": 1629.669921875, "completions/mean_terminated_length": 1184.350830078125, "completions/min_length": 72.0, "completions/min_terminated_length": 72.0, "epoch": 0.04608688230775796, "frac_reward_zero_std": 0.625, "grad_norm": 0.06083833926021826, "learning_rate": 4.573378839590443e-07, "loss": 0.0208, "num_tokens": 130359138.0, "reward": 0.125, "reward_std": 0.17249077558517456, "rewards/accuracy_reward/mean": 0.125, "rewards/accuracy_reward/std": 0.3310423493385315, "step": 135 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.548828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2028.0, "completions/mean_length": 1708.466796875, "completions/mean_terminated_length": 1295.4415283203125, "completions/min_length": 52.0, "completions/min_terminated_length": 52.0, "epoch": 0.04642826662114876, "frac_reward_zero_std": 0.65625, "grad_norm": 0.05601725402255147, "learning_rate": 4.6075085324232084e-07, "loss": 0.0254, "num_tokens": 131312481.0, "reward": 0.10546875, "reward_std": 0.14787328243255615, "rewards/accuracy_reward/mean": 0.1088709682226181, "rewards/accuracy_reward/std": 0.31179171800613403, "step": 136 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.666015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2038.0, "completions/mean_length": 1736.83984375, "completions/mean_terminated_length": 1116.9532470703125, "completions/min_length": 31.0, "completions/min_terminated_length": 31.0, "epoch": 0.04676965093453956, "frac_reward_zero_std": 0.6875, "grad_norm": 0.05588237673953811, "learning_rate": 4.641638225255973e-07, "loss": 0.0052, "num_tokens": 132288063.0, "reward": 0.060546875, "reward_std": 0.11057336628437042, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 137 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.529296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2035.0, "completions/mean_length": 1639.3671875, "completions/mean_terminated_length": 1179.8673095703125, "completions/min_length": 202.0, "completions/min_terminated_length": 202.0, "epoch": 0.04711103524793036, "frac_reward_zero_std": 0.8125, "grad_norm": 0.038698838885418696, "learning_rate": 4.675767918088737e-07, "loss": 0.0134, "num_tokens": 133198779.0, "reward": 0.041015625, "reward_std": 0.07250870764255524, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 138 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.630859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2041.0, "completions/mean_length": 1722.880859375, "completions/mean_terminated_length": 1167.25390625, "completions/min_length": 283.0, "completions/min_terminated_length": 283.0, "epoch": 0.04745241956132116, "frac_reward_zero_std": 0.75, "grad_norm": 0.04487318351480335, "learning_rate": 4.709897610921502e-07, "loss": 0.013, "num_tokens": 134162318.0, "reward": 0.05859375, "reward_std": 0.08683578670024872, "rewards/accuracy_reward/mean": 0.05859375, "rewards/accuracy_reward/std": 0.23509246110916138, "step": 139 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.63671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2040.0, "completions/mean_length": 1734.521484375, "completions/mean_terminated_length": 1185.0914306640625, "completions/min_length": 170.0, "completions/min_terminated_length": 170.0, "epoch": 0.04779380387471196, "frac_reward_zero_std": 0.75, "grad_norm": 0.05003938325459597, "learning_rate": 4.744027303754266e-07, "loss": 0.023, "num_tokens": 135121817.0, "reward": 0.0703125, "reward_std": 0.09244654327630997, "rewards/accuracy_reward/mean": 0.07258064299821854, "rewards/accuracy_reward/std": 0.25970885157585144, "step": 140 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2032.0, "completions/mean_length": 1661.90234375, "completions/mean_terminated_length": 1132.8055419921875, "completions/min_length": 95.0, "completions/min_terminated_length": 95.0, "epoch": 0.04813518818810276, "frac_reward_zero_std": 0.59375, "grad_norm": 0.05694080088244041, "learning_rate": 4.778156996587031e-07, "loss": 0.0299, "num_tokens": 136050215.0, "reward": 0.087890625, "reward_std": 0.1455163061618805, "rewards/accuracy_reward/mean": 0.087890625, "rewards/accuracy_reward/std": 0.2834126651287079, "step": 141 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.59765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1720.783203125, "completions/mean_terminated_length": 1234.7232666015625, "completions/min_length": 368.0, "completions/min_terminated_length": 368.0, "epoch": 0.04847657250149356, "frac_reward_zero_std": 0.71875, "grad_norm": 0.0534909894123429, "learning_rate": 4.812286689419795e-07, "loss": 0.0276, "num_tokens": 137014536.0, "reward": 0.0703125, "reward_std": 0.10674113780260086, "rewards/accuracy_reward/mean": 0.0703125, "rewards/accuracy_reward/std": 0.25592297315597534, "step": 142 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.58984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2045.0, "completions/mean_length": 1689.265625, "completions/mean_terminated_length": 1173.3714599609375, "completions/min_length": 28.0, "completions/min_terminated_length": 28.0, "epoch": 0.04881795681488436, "frac_reward_zero_std": 0.78125, "grad_norm": 0.04011415363157658, "learning_rate": 4.84641638225256e-07, "loss": 0.0282, "num_tokens": 137952640.0, "reward": 0.056640625, "reward_std": 0.07686128467321396, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 143 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.55078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2021.0, "completions/mean_length": 1676.654296875, "completions/mean_terminated_length": 1221.5260009765625, "completions/min_length": 45.0, "completions/min_terminated_length": 45.0, "epoch": 0.04915934112827516, "frac_reward_zero_std": 0.75, "grad_norm": 0.04582785148181721, "learning_rate": 4.880546075085323e-07, "loss": 0.0224, "num_tokens": 138890623.0, "reward": 0.10546875, "reward_std": 0.1024416983127594, "rewards/accuracy_reward/mean": 0.10546875, "rewards/accuracy_reward/std": 0.3074568510055542, "step": 144 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.427734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2021.0, "completions/mean_length": 1472.955078125, "completions/mean_terminated_length": 1043.143310546875, "completions/min_length": 90.0, "completions/min_terminated_length": 90.0, "epoch": 0.049500725441665956, "frac_reward_zero_std": 0.625, "grad_norm": 0.07300036922061065, "learning_rate": 4.914675767918088e-07, "loss": 0.0198, "num_tokens": 139715080.0, "reward": 0.1328125, "reward_std": 0.15780803561210632, "rewards/accuracy_reward/mean": 0.1328125, "rewards/accuracy_reward/std": 0.33970388770103455, "step": 145 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.5390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2037.0, "completions/mean_length": 1656.533203125, "completions/mean_terminated_length": 1198.716064453125, "completions/min_length": 22.0, "completions/min_terminated_length": 22.0, "epoch": 0.049842109755056756, "frac_reward_zero_std": 0.6875, "grad_norm": 0.04804002639321947, "learning_rate": 4.948805460750853e-07, "loss": 0.011, "num_tokens": 140637401.0, "reward": 0.091796875, "reward_std": 0.12312431633472443, "rewards/accuracy_reward/mean": 0.091796875, "rewards/accuracy_reward/std": 0.289021372795105, "step": 146 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.5078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.0, "completions/mean_length": 1589.837890625, "completions/mean_terminated_length": 1117.1309814453125, "completions/min_length": 84.0, "completions/min_terminated_length": 84.0, "epoch": 0.050183494068447555, "frac_reward_zero_std": 0.71875, "grad_norm": 0.0479645443070329, "learning_rate": 4.982935153583617e-07, "loss": 0.0256, "num_tokens": 141530518.0, "reward": 0.0703125, "reward_std": 0.10881631821393967, "rewards/accuracy_reward/mean": 0.07258064299821854, "rewards/accuracy_reward/std": 0.25970885157585144, "step": 147 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.52734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2039.0, "completions/mean_length": 1681.123046875, "completions/mean_terminated_length": 1271.7974853515625, "completions/min_length": 252.0, "completions/min_terminated_length": 252.0, "epoch": 0.050524878381838355, "frac_reward_zero_std": 0.625, "grad_norm": 0.05437882003712324, "learning_rate": 5.017064846416383e-07, "loss": 0.0127, "num_tokens": 142470005.0, "reward": 0.1328125, "reward_std": 0.14288797974586487, "rewards/accuracy_reward/mean": 0.1328125, "rewards/accuracy_reward/std": 0.33970388770103455, "step": 148 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.607421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.0, "completions/mean_length": 1681.205078125, "completions/mean_terminated_length": 1113.6766357421875, "completions/min_length": 13.0, "completions/min_terminated_length": 13.0, "epoch": 0.050866262695229154, "frac_reward_zero_std": 0.71875, "grad_norm": 0.05242939089020235, "learning_rate": 5.051194539249146e-07, "loss": 0.0275, "num_tokens": 143408862.0, "reward": 0.099609375, "reward_std": 0.09644746035337448, "rewards/accuracy_reward/mean": 0.099609375, "rewards/accuracy_reward/std": 0.29977133870124817, "step": 149 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.57421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2045.0, "completions/mean_length": 1638.263671875, "completions/mean_terminated_length": 1085.6834716796875, "completions/min_length": 118.0, "completions/min_terminated_length": 118.0, "epoch": 0.051207647008619954, "frac_reward_zero_std": 0.6875, "grad_norm": 0.04690025492945931, "learning_rate": 5.085324232081911e-07, "loss": 0.0106, "num_tokens": 144322853.0, "reward": 0.076171875, "reward_std": 0.10546942055225372, "rewards/accuracy_reward/mean": 0.076171875, "rewards/accuracy_reward/std": 0.26553234457969666, "step": 150 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.587890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2042.0, "completions/mean_length": 1712.775390625, "completions/mean_terminated_length": 1234.8199462890625, "completions/min_length": 42.0, "completions/min_terminated_length": 42.0, "epoch": 0.051549031322010753, "frac_reward_zero_std": 0.71875, "grad_norm": 0.05088571292453204, "learning_rate": 5.119453924914675e-07, "loss": 0.014, "num_tokens": 145280562.0, "reward": 0.0546875, "reward_std": 0.10469119250774384, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 151 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.5234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1651.888671875, "completions/mean_terminated_length": 1216.913818359375, "completions/min_length": 17.0, "completions/min_terminated_length": 17.0, "epoch": 0.05189041563540155, "frac_reward_zero_std": 0.6875, "grad_norm": 0.05625865566795103, "learning_rate": 5.15358361774744e-07, "loss": 0.0049, "num_tokens": 146210665.0, "reward": 0.146484375, "reward_std": 0.12141455709934235, "rewards/accuracy_reward/mean": 0.146484375, "rewards/accuracy_reward/std": 0.35393697023391724, "step": 152 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.638671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1757.263671875, "completions/mean_terminated_length": 1243.3675537109375, "completions/min_length": 200.0, "completions/min_terminated_length": 200.0, "epoch": 0.05223179994879235, "frac_reward_zero_std": 0.78125, "grad_norm": 0.03961015669745025, "learning_rate": 5.187713310580204e-07, "loss": 0.0176, "num_tokens": 147184944.0, "reward": 0.044921875, "reward_std": 0.08483455330133438, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 153 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.48828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2043.0, "completions/mean_length": 1585.193359375, "completions/mean_terminated_length": 1143.583984375, "completions/min_length": 47.0, "completions/min_terminated_length": 47.0, "epoch": 0.05257318426218315, "frac_reward_zero_std": 0.71875, "grad_norm": 0.05207746909601062, "learning_rate": 5.221843003412969e-07, "loss": 0.006, "num_tokens": 148070723.0, "reward": 0.111328125, "reward_std": 0.1081790179014206, "rewards/accuracy_reward/mean": 0.111328125, "rewards/accuracy_reward/std": 0.31484565138816833, "step": 154 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2034.0, "completions/mean_length": 1613.388671875, "completions/mean_terminated_length": 1150.7379150390625, "completions/min_length": 180.0, "completions/min_terminated_length": 180.0, "epoch": 0.05291456857557395, "frac_reward_zero_std": 0.6875, "grad_norm": 0.05803034337918793, "learning_rate": 5.255972696245734e-07, "loss": 0.0137, "num_tokens": 148964106.0, "reward": 0.115234375, "reward_std": 0.13362500071525574, "rewards/accuracy_reward/mean": 0.11895161122083664, "rewards/accuracy_reward/std": 0.3240583837032318, "step": 155 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.513671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2048.0, "completions/mean_length": 1575.779296875, "completions/mean_terminated_length": 1077.0079345703125, "completions/min_length": 18.0, "completions/min_terminated_length": 18.0, "epoch": 0.05325595288896475, "frac_reward_zero_std": 0.59375, "grad_norm": 0.0683698311192887, "learning_rate": 5.290102389078498e-07, "loss": 0.0127, "num_tokens": 149843913.0, "reward": 0.10546875, "reward_std": 0.15122081339359283, "rewards/accuracy_reward/mean": 0.10546875, "rewards/accuracy_reward/std": 0.3074568510055542, "step": 156 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.54296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2003.0, "completions/mean_length": 1590.623046875, "completions/mean_terminated_length": 1047.24365234375, "completions/min_length": 121.0, "completions/min_terminated_length": 121.0, "epoch": 0.05359733720235555, "frac_reward_zero_std": 0.65625, "grad_norm": 0.05589296281262381, "learning_rate": 5.324232081911263e-07, "loss": 0.0187, "num_tokens": 150728152.0, "reward": 0.1171875, "reward_std": 0.13104984164237976, "rewards/accuracy_reward/mean": 0.1171875, "rewards/accuracy_reward/std": 0.32195815443992615, "step": 157 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.52734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2017.0, "completions/mean_length": 1591.783203125, "completions/mean_terminated_length": 1082.781005859375, "completions/min_length": 204.0, "completions/min_terminated_length": 204.0, "epoch": 0.05393872151574635, "frac_reward_zero_std": 0.6875, "grad_norm": 0.052967163872803126, "learning_rate": 5.358361774744027e-07, "loss": 0.0251, "num_tokens": 151623609.0, "reward": 0.056640625, "reward_std": 0.08907270431518555, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 158 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.51171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1602.890625, "completions/mean_terminated_length": 1136.416015625, "completions/min_length": 80.0, "completions/min_terminated_length": 80.0, "epoch": 0.05428010582913715, "frac_reward_zero_std": 0.65625, "grad_norm": 0.05292779656327808, "learning_rate": 5.392491467576792e-07, "loss": 0.0165, "num_tokens": 152524513.0, "reward": 0.087890625, "reward_std": 0.12337085604667664, "rewards/accuracy_reward/mean": 0.09072580933570862, "rewards/accuracy_reward/std": 0.2875087857246399, "step": 159 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.517578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2040.0, "completions/mean_length": 1602.990234375, "completions/mean_terminated_length": 1125.5506591796875, "completions/min_length": 25.0, "completions/min_terminated_length": 25.0, "epoch": 0.05462149014252795, "frac_reward_zero_std": 0.625, "grad_norm": 0.05872180443805259, "learning_rate": 5.426621160409555e-07, "loss": 0.0063, "num_tokens": 153422156.0, "reward": 0.169921875, "reward_std": 0.16396436095237732, "rewards/accuracy_reward/mean": 0.169921875, "rewards/accuracy_reward/std": 0.3759314715862274, "step": 160 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1602.109375, "completions/mean_terminated_length": 1232.6571044921875, "completions/min_length": 34.0, "completions/min_terminated_length": 34.0, "epoch": 0.05496287445591875, "frac_reward_zero_std": 0.6875, "grad_norm": 0.06237912695561149, "learning_rate": 5.46075085324232e-07, "loss": 0.0148, "num_tokens": 154324212.0, "reward": 0.134765625, "reward_std": 0.13336601853370667, "rewards/accuracy_reward/mean": 0.134765625, "rewards/accuracy_reward/std": 0.3418070077896118, "step": 161 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.51953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2025.0, "completions/mean_length": 1590.720703125, "completions/mean_terminated_length": 1096.26416015625, "completions/min_length": 46.0, "completions/min_terminated_length": 46.0, "epoch": 0.05530425876930955, "frac_reward_zero_std": 0.71875, "grad_norm": 0.04513467933459144, "learning_rate": 5.494880546075085e-07, "loss": 0.0185, "num_tokens": 155215333.0, "reward": 0.111328125, "reward_std": 0.1183076947927475, "rewards/accuracy_reward/mean": 0.111328125, "rewards/accuracy_reward/std": 0.31484565138816833, "step": 162 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.44921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1538.525390625, "completions/mean_terminated_length": 1122.9964599609375, "completions/min_length": 83.0, "completions/min_terminated_length": 83.0, "epoch": 0.05564564308270035, "frac_reward_zero_std": 0.65625, "grad_norm": 0.05635666183435418, "learning_rate": 5.52901023890785e-07, "loss": 0.028, "num_tokens": 156079586.0, "reward": 0.126953125, "reward_std": 0.13561506569385529, "rewards/accuracy_reward/mean": 0.126953125, "rewards/accuracy_reward/std": 0.33324605226516724, "step": 163 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.4453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2048.0, "completions/mean_length": 1500.015625, "completions/mean_terminated_length": 1060.08447265625, "completions/min_length": 25.0, "completions/min_terminated_length": 25.0, "epoch": 0.05598702739609115, "frac_reward_zero_std": 0.71875, "grad_norm": 0.056823796809735344, "learning_rate": 5.563139931740614e-07, "loss": 0.0213, "num_tokens": 156925530.0, "reward": 0.13671875, "reward_std": 0.09655256569385529, "rewards/accuracy_reward/mean": 0.13671875, "rewards/accuracy_reward/std": 0.3438861668109894, "step": 164 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.576171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2048.0, "completions/mean_length": 1651.23046875, "completions/mean_terminated_length": 1111.84326171875, "completions/min_length": 61.0, "completions/min_terminated_length": 61.0, "epoch": 0.05632841170948195, "frac_reward_zero_std": 0.71875, "grad_norm": 0.04406791313036526, "learning_rate": 5.597269624573379e-07, "loss": 0.0261, "num_tokens": 157846880.0, "reward": 0.08203125, "reward_std": 0.11667752265930176, "rewards/accuracy_reward/mean": 0.08203125, "rewards/accuracy_reward/std": 0.2746807038784027, "step": 165 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.345703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2035.0, "completions/mean_length": 1348.568359375, "completions/mean_terminated_length": 979.0179443359375, "completions/min_length": 28.0, "completions/min_terminated_length": 28.0, "epoch": 0.05666979602287275, "frac_reward_zero_std": 0.59375, "grad_norm": 0.062472373957335216, "learning_rate": 5.631399317406143e-07, "loss": 0.0255, "num_tokens": 158614419.0, "reward": 0.146484375, "reward_std": 0.18024902045726776, "rewards/accuracy_reward/mean": 0.146484375, "rewards/accuracy_reward/std": 0.35393697023391724, "step": 166 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.427734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2026.0, "completions/mean_length": 1526.673828125, "completions/mean_terminated_length": 1137.013671875, "completions/min_length": 247.0, "completions/min_terminated_length": 247.0, "epoch": 0.057011180336263546, "frac_reward_zero_std": 0.625, "grad_norm": 0.05488752739517643, "learning_rate": 5.665529010238907e-07, "loss": 0.0176, "num_tokens": 159466684.0, "reward": 0.125, "reward_std": 0.13437044620513916, "rewards/accuracy_reward/mean": 0.125, "rewards/accuracy_reward/std": 0.3310423493385315, "step": 167 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2001.0, "completions/mean_length": 1626.142578125, "completions/mean_terminated_length": 1048.4305419921875, "completions/min_length": 30.0, "completions/min_terminated_length": 30.0, "epoch": 0.057352564649654346, "frac_reward_zero_std": 0.75, "grad_norm": 0.05434770049927932, "learning_rate": 5.699658703071673e-07, "loss": 0.0121, "num_tokens": 160385877.0, "reward": 0.0703125, "reward_std": 0.09566686302423477, "rewards/accuracy_reward/mean": 0.0703125, "rewards/accuracy_reward/std": 0.25592297315597534, "step": 168 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.4296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2042.0, "completions/mean_length": 1486.888671875, "completions/mean_terminated_length": 1064.133544921875, "completions/min_length": 80.0, "completions/min_terminated_length": 80.0, "epoch": 0.057693948963045145, "frac_reward_zero_std": 0.625, "grad_norm": 0.07295831048210671, "learning_rate": 5.733788395904437e-07, "loss": 0.0144, "num_tokens": 161220524.0, "reward": 0.13671875, "reward_std": 0.1389889121055603, "rewards/accuracy_reward/mean": 0.13671875, "rewards/accuracy_reward/std": 0.3438861668109894, "step": 169 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.55859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2040.0, "completions/mean_length": 1558.830078125, "completions/mean_terminated_length": 939.7920532226562, "completions/min_length": 35.0, "completions/min_terminated_length": 35.0, "epoch": 0.058035333276435945, "frac_reward_zero_std": 0.75, "grad_norm": 0.05151259468658889, "learning_rate": 5.767918088737202e-07, "loss": 0.0026, "num_tokens": 162097525.0, "reward": 0.0390625, "reward_std": 0.08917921781539917, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 170 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.46875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2038.0, "completions/mean_length": 1555.212890625, "completions/mean_terminated_length": 1120.4007568359375, "completions/min_length": 14.0, "completions/min_terminated_length": 14.0, "epoch": 0.058376717589826745, "frac_reward_zero_std": 0.65625, "grad_norm": 0.0532030445654481, "learning_rate": 5.802047781569965e-07, "loss": 0.0132, "num_tokens": 162971298.0, "reward": 0.1328125, "reward_std": 0.1513998806476593, "rewards/accuracy_reward/mean": 0.1328125, "rewards/accuracy_reward/std": 0.33970388770103455, "step": 171 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.42578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2035.0, "completions/mean_length": 1442.48046875, "completions/mean_terminated_length": 993.48974609375, "completions/min_length": 25.0, "completions/min_terminated_length": 25.0, "epoch": 0.058718101903217544, "frac_reward_zero_std": 0.6875, "grad_norm": 0.05577109989592277, "learning_rate": 5.83617747440273e-07, "loss": 0.0146, "num_tokens": 163795480.0, "reward": 0.099609375, "reward_std": 0.1336861550807953, "rewards/accuracy_reward/mean": 0.099609375, "rewards/accuracy_reward/std": 0.29977133870124817, "step": 172 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.447265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2041.0, "completions/mean_length": 1538.205078125, "completions/mean_terminated_length": 1125.685546875, "completions/min_length": 170.0, "completions/min_terminated_length": 170.0, "epoch": 0.059059486216608344, "frac_reward_zero_std": 0.625, "grad_norm": 0.049739956106912495, "learning_rate": 5.870307167235494e-07, "loss": 0.0068, "num_tokens": 164658753.0, "reward": 0.1015625, "reward_std": 0.14584887027740479, "rewards/accuracy_reward/mean": 0.1015625, "rewards/accuracy_reward/std": 0.30236753821372986, "step": 173 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.388671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2044.0, "completions/mean_length": 1346.8046875, "completions/mean_terminated_length": 900.9967651367188, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.05940087052999914, "frac_reward_zero_std": 0.59375, "grad_norm": 0.07032360980457225, "learning_rate": 5.904436860068259e-07, "loss": 0.0008, "num_tokens": 165428333.0, "reward": 0.1484375, "reward_std": 0.1553596556186676, "rewards/accuracy_reward/mean": 0.15322580933570862, "rewards/accuracy_reward/std": 0.36056873202323914, "step": 174 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.36328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2048.0, "completions/mean_length": 1320.203125, "completions/mean_terminated_length": 904.95703125, "completions/min_length": 47.0, "completions/min_terminated_length": 47.0, "epoch": 0.05974225484338994, "frac_reward_zero_std": 0.6875, "grad_norm": 0.06093328097141317, "learning_rate": 5.938566552901024e-07, "loss": 0.0237, "num_tokens": 166182021.0, "reward": 0.10546875, "reward_std": 0.11872635036706924, "rewards/accuracy_reward/mean": 0.10546875, "rewards/accuracy_reward/std": 0.3074568510055542, "step": 175 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.458984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2039.0, "completions/mean_length": 1458.333984375, "completions/mean_terminated_length": 958.0758056640625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.06008363915678075, "frac_reward_zero_std": 0.71875, "grad_norm": 0.058548190775278085, "learning_rate": 5.972696245733788e-07, "loss": 0.0173, "num_tokens": 167008544.0, "reward": 0.091796875, "reward_std": 0.11953167617321014, "rewards/accuracy_reward/mean": 0.0947580635547638, "rewards/accuracy_reward/std": 0.29317617416381836, "step": 176 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.443359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2037.0, "completions/mean_length": 1467.755859375, "completions/mean_terminated_length": 1005.5964965820312, "completions/min_length": 41.0, "completions/min_terminated_length": 41.0, "epoch": 0.06042502347017155, "frac_reward_zero_std": 0.59375, "grad_norm": 0.06136394213803249, "learning_rate": 6.006825938566553e-07, "loss": 0.0237, "num_tokens": 167837507.0, "reward": 0.125, "reward_std": 0.14988242089748383, "rewards/accuracy_reward/mean": 0.125, "rewards/accuracy_reward/std": 0.3310423493385315, "step": 177 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.427734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2044.0, "completions/mean_length": 1411.6171875, "completions/mean_terminated_length": 935.9590454101562, "completions/min_length": 20.0, "completions/min_terminated_length": 20.0, "epoch": 0.06076640778356235, "frac_reward_zero_std": 0.6875, "grad_norm": 0.05365700955680286, "learning_rate": 6.040955631399317e-07, "loss": 0.0347, "num_tokens": 168639391.0, "reward": 0.130859375, "reward_std": 0.13703955709934235, "rewards/accuracy_reward/mean": 0.130859375, "rewards/accuracy_reward/std": 0.33757632970809937, "step": 178 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.40234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1390.310546875, "completions/mean_terminated_length": 947.5523071289062, "completions/min_length": 23.0, "completions/min_terminated_length": 23.0, "epoch": 0.06110779209695315, "frac_reward_zero_std": 0.78125, "grad_norm": 0.05777821883833454, "learning_rate": 6.075085324232082e-07, "loss": 0.03, "num_tokens": 169425358.0, "reward": 0.09765625, "reward_std": 0.09079696238040924, "rewards/accuracy_reward/mean": 0.10080645233392715, "rewards/accuracy_reward/std": 0.30137622356414795, "step": 179 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2040.0, "completions/mean_length": 1482.677734375, "completions/mean_terminated_length": 951.617431640625, "completions/min_length": 10.0, "completions/min_terminated_length": 10.0, "epoch": 0.06144917641034395, "frac_reward_zero_std": 0.78125, "grad_norm": 0.049112555183154175, "learning_rate": 6.109215017064846e-07, "loss": 0.0165, "num_tokens": 170273577.0, "reward": 0.052734375, "reward_std": 0.08887359499931335, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 180 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.271484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2027.0, "completions/mean_length": 1139.21875, "completions/mean_terminated_length": 800.5576782226562, "completions/min_length": 42.0, "completions/min_terminated_length": 42.0, "epoch": 0.06179056072373475, "frac_reward_zero_std": 0.8125, "grad_norm": 0.06202726901337909, "learning_rate": 6.143344709897611e-07, "loss": 0.0071, "num_tokens": 170938505.0, "reward": 0.123046875, "reward_std": 0.04973640665411949, "rewards/accuracy_reward/mean": 0.123046875, "rewards/accuracy_reward/std": 0.32881227135658264, "step": 181 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.40234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2037.0, "completions/mean_length": 1410.91796875, "completions/mean_terminated_length": 982.03271484375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.06213194503712555, "frac_reward_zero_std": 0.78125, "grad_norm": 0.04369301292203226, "learning_rate": 6.177474402730375e-07, "loss": 0.0071, "num_tokens": 171739055.0, "reward": 0.048828125, "reward_std": 0.08207826316356659, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 182 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.263671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2042.0, "completions/mean_length": 1174.68359375, "completions/mean_terminated_length": 861.95751953125, "completions/min_length": 11.0, "completions/min_terminated_length": 11.0, "epoch": 0.062473329350516346, "frac_reward_zero_std": 0.8125, "grad_norm": 0.04384792156315624, "learning_rate": 6.21160409556314e-07, "loss": 0.0163, "num_tokens": 172416845.0, "reward": 0.083984375, "reward_std": 0.08164606243371964, "rewards/accuracy_reward/mean": 0.083984375, "rewards/accuracy_reward/std": 0.2776356339454651, "step": 183 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.37109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2038.0, "completions/mean_length": 1412.220703125, "completions/mean_terminated_length": 1037.0714111328125, "completions/min_length": 191.0, "completions/min_terminated_length": 191.0, "epoch": 0.06281471366390715, "frac_reward_zero_std": 0.625, "grad_norm": 0.0845404281614352, "learning_rate": 6.245733788395904e-07, "loss": 0.0095, "num_tokens": 173215886.0, "reward": 0.12890625, "reward_std": 0.13614104688167572, "rewards/accuracy_reward/mean": 0.12890625, "rewards/accuracy_reward/std": 0.33542385697364807, "step": 184 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.388671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2042.0, "completions/mean_length": 1381.515625, "completions/mean_terminated_length": 957.9520874023438, "completions/min_length": 25.0, "completions/min_terminated_length": 25.0, "epoch": 0.06315609797729795, "frac_reward_zero_std": 0.71875, "grad_norm": 0.04946703476560849, "learning_rate": 6.279863481228669e-07, "loss": 0.0121, "num_tokens": 174002726.0, "reward": 0.125, "reward_std": 0.1180124506354332, "rewards/accuracy_reward/mean": 0.12903225421905518, "rewards/accuracy_reward/std": 0.33557409048080444, "step": 185 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.408203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2044.0, "completions/mean_length": 1410.26171875, "completions/mean_terminated_length": 970.5050048828125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.06349748229068874, "frac_reward_zero_std": 0.4375, "grad_norm": 0.07915382776717433, "learning_rate": 6.313993174061433e-07, "loss": 0.0067, "num_tokens": 174805436.0, "reward": 0.220703125, "reward_std": 0.21836557984352112, "rewards/accuracy_reward/mean": 0.220703125, "rewards/accuracy_reward/std": 0.4151262938976288, "step": 186 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.212890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2040.0, "completions/mean_length": 1071.439453125, "completions/mean_terminated_length": 807.3076782226562, "completions/min_length": 45.0, "completions/min_terminated_length": 45.0, "epoch": 0.06383886660407954, "frac_reward_zero_std": 0.5, "grad_norm": 0.09553243134690065, "learning_rate": 6.348122866894197e-07, "loss": 0.0328, "num_tokens": 175428957.0, "reward": 0.185546875, "reward_std": 0.1863960325717926, "rewards/accuracy_reward/mean": 0.185546875, "rewards/accuracy_reward/std": 0.38912075757980347, "step": 187 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.33203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2039.0, "completions/mean_length": 1284.94140625, "completions/mean_terminated_length": 905.643310546875, "completions/min_length": 24.0, "completions/min_terminated_length": 24.0, "epoch": 0.06418025091747034, "frac_reward_zero_std": 0.75, "grad_norm": 0.059261370227590045, "learning_rate": 6.382252559726961e-07, "loss": 0.0156, "num_tokens": 176164719.0, "reward": 0.1015625, "reward_std": 0.10013246536254883, "rewards/accuracy_reward/mean": 0.1015625, "rewards/accuracy_reward/std": 0.30236753821372986, "step": 188 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.33984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2023.0, "completions/mean_length": 1222.591796875, "completions/mean_terminated_length": 797.6775512695312, "completions/min_length": 50.0, "completions/min_terminated_length": 50.0, "epoch": 0.06452163523086114, "frac_reward_zero_std": 0.75, "grad_norm": 0.0598591479349171, "learning_rate": 6.416382252559727e-07, "loss": 0.0076, "num_tokens": 176867214.0, "reward": 0.08203125, "reward_std": 0.09244653582572937, "rewards/accuracy_reward/mean": 0.08203125, "rewards/accuracy_reward/std": 0.2746807038784027, "step": 189 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.365234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2039.0, "completions/mean_length": 1270.1171875, "completions/mean_terminated_length": 822.535400390625, "completions/min_length": 27.0, "completions/min_terminated_length": 27.0, "epoch": 0.06486301954425194, "frac_reward_zero_std": 0.5625, "grad_norm": 0.06982183160504854, "learning_rate": 6.450511945392492e-07, "loss": 0.0052, "num_tokens": 177590010.0, "reward": 0.091796875, "reward_std": 0.1510140299797058, "rewards/accuracy_reward/mean": 0.091796875, "rewards/accuracy_reward/std": 0.289021372795105, "step": 190 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.388671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2048.0, "completions/mean_length": 1385.404296875, "completions/mean_terminated_length": 964.1373901367188, "completions/min_length": 49.0, "completions/min_terminated_length": 49.0, "epoch": 0.06520440385764274, "frac_reward_zero_std": 0.46875, "grad_norm": 0.07297420217316669, "learning_rate": 6.484641638225256e-07, "loss": 0.0418, "num_tokens": 178380953.0, "reward": 0.162109375, "reward_std": 0.21253521740436554, "rewards/accuracy_reward/mean": 0.162109375, "rewards/accuracy_reward/std": 0.3689115643501282, "step": 191 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.33984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2018.0, "completions/mean_length": 1354.4609375, "completions/mean_terminated_length": 997.5621337890625, "completions/min_length": 19.0, "completions/min_terminated_length": 19.0, "epoch": 0.06554578817103354, "frac_reward_zero_std": 0.65625, "grad_norm": 0.07188494202544887, "learning_rate": 6.51877133105802e-07, "loss": -0.0046, "num_tokens": 179149813.0, "reward": 0.064453125, "reward_std": 0.12636467814445496, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 192 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.318359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2032.0, "completions/mean_length": 1265.36328125, "completions/mean_terminated_length": 899.8338623046875, "completions/min_length": 25.0, "completions/min_terminated_length": 25.0, "epoch": 0.06588717248442434, "frac_reward_zero_std": 0.65625, "grad_norm": 0.06838047937663966, "learning_rate": 6.552901023890784e-07, "loss": 0.007, "num_tokens": 179881423.0, "reward": 0.109375, "reward_std": 0.1336916983127594, "rewards/accuracy_reward/mean": 0.109375, "rewards/accuracy_reward/std": 0.31241437792778015, "step": 193 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2044.0, "completions/mean_length": 1234.23828125, "completions/mean_terminated_length": 890.6500244140625, "completions/min_length": 17.0, "completions/min_terminated_length": 17.0, "epoch": 0.06622855679781514, "frac_reward_zero_std": 0.625, "grad_norm": 0.061814204134840234, "learning_rate": 6.587030716723549e-07, "loss": 0.0132, "num_tokens": 180591465.0, "reward": 0.123046875, "reward_std": 0.13055095076560974, "rewards/accuracy_reward/mean": 0.1270161271095276, "rewards/accuracy_reward/std": 0.33332720398902893, "step": 194 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.306640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2035.0, "completions/mean_length": 1238.390625, "completions/mean_terminated_length": 880.3380126953125, "completions/min_length": 11.0, "completions/min_terminated_length": 11.0, "epoch": 0.06656994111120594, "frac_reward_zero_std": 0.59375, "grad_norm": 0.06388538599344679, "learning_rate": 6.621160409556313e-07, "loss": 0.0096, "num_tokens": 181298961.0, "reward": 0.12109375, "reward_std": 0.1446847766637802, "rewards/accuracy_reward/mean": 0.12109375, "rewards/accuracy_reward/std": 0.3265552520751953, "step": 195 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.251953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2043.0, "completions/mean_length": 1193.0, "completions/mean_terminated_length": 905.0234985351562, "completions/min_length": 25.0, "completions/min_terminated_length": 25.0, "epoch": 0.06691132542459674, "frac_reward_zero_std": 0.65625, "grad_norm": 0.06061534148759774, "learning_rate": 6.655290102389079e-07, "loss": 0.0095, "num_tokens": 181985825.0, "reward": 0.1328125, "reward_std": 0.14849212765693665, "rewards/accuracy_reward/mean": 0.1328125, "rewards/accuracy_reward/std": 0.33970388770103455, "step": 196 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.34765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2022.0, "completions/mean_length": 1370.814453125, "completions/mean_terminated_length": 1009.919189453125, "completions/min_length": 45.0, "completions/min_terminated_length": 45.0, "epoch": 0.06725270973798754, "frac_reward_zero_std": 0.71875, "grad_norm": 0.05111044258348449, "learning_rate": 6.689419795221843e-07, "loss": 0.0264, "num_tokens": 182762882.0, "reward": 0.123046875, "reward_std": 0.10535639524459839, "rewards/accuracy_reward/mean": 0.123046875, "rewards/accuracy_reward/std": 0.32881227135658264, "step": 197 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.2578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2048.0, "completions/mean_length": 1217.041015625, "completions/mean_terminated_length": 928.3921508789062, "completions/min_length": 25.0, "completions/min_terminated_length": 25.0, "epoch": 0.06759409405137834, "frac_reward_zero_std": 0.6875, "grad_norm": 0.06303566480342586, "learning_rate": 6.723549488054607e-07, "loss": 0.0126, "num_tokens": 183457335.0, "reward": 0.130859375, "reward_std": 0.12752877175807953, "rewards/accuracy_reward/mean": 0.130859375, "rewards/accuracy_reward/std": 0.33757632970809937, "step": 198 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.318359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2028.0, "completions/mean_length": 1354.201171875, "completions/mean_terminated_length": 1030.163330078125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.06793547836476914, "frac_reward_zero_std": 0.65625, "grad_norm": 0.054702744240987866, "learning_rate": 6.757679180887371e-07, "loss": 0.0171, "num_tokens": 184227998.0, "reward": 0.1640625, "reward_std": 0.12148028612136841, "rewards/accuracy_reward/mean": 0.1640625, "rewards/accuracy_reward/std": 0.37069445848464966, "step": 199 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.369140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2040.0, "completions/mean_length": 1379.123046875, "completions/mean_terminated_length": 987.73681640625, "completions/min_length": 15.0, "completions/min_terminated_length": 15.0, "epoch": 0.06827686267815994, "frac_reward_zero_std": 0.6875, "grad_norm": 0.05259435310981436, "learning_rate": 6.791808873720136e-07, "loss": 0.0147, "num_tokens": 185001821.0, "reward": 0.125, "reward_std": 0.131615549325943, "rewards/accuracy_reward/mean": 0.125, "rewards/accuracy_reward/std": 0.3310423493385315, "step": 200 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.33984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2042.0, "completions/mean_length": 1317.5546875, "completions/mean_terminated_length": 941.526611328125, "completions/min_length": 25.0, "completions/min_terminated_length": 25.0, "epoch": 0.06861824699155074, "frac_reward_zero_std": 0.75, "grad_norm": 0.05284857738543805, "learning_rate": 6.825938566552901e-07, "loss": -0.0033, "num_tokens": 185753929.0, "reward": 0.07421875, "reward_std": 0.10067910701036453, "rewards/accuracy_reward/mean": 0.07421875, "rewards/accuracy_reward/std": 0.2623828947544098, "step": 201 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.43359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2029.0, "completions/mean_length": 1418.537109375, "completions/mean_terminated_length": 936.6724243164062, "completions/min_length": 22.0, "completions/min_terminated_length": 22.0, "epoch": 0.06895963130494154, "frac_reward_zero_std": 0.53125, "grad_norm": 0.07815481352799261, "learning_rate": 6.860068259385665e-07, "loss": 0.0351, "num_tokens": 186564364.0, "reward": 0.158203125, "reward_std": 0.17998173832893372, "rewards/accuracy_reward/mean": 0.158203125, "rewards/accuracy_reward/std": 0.36528825759887695, "step": 202 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.36328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2039.0, "completions/mean_length": 1348.0078125, "completions/mean_terminated_length": 948.625732421875, "completions/min_length": 18.0, "completions/min_terminated_length": 18.0, "epoch": 0.06930101561833234, "frac_reward_zero_std": 0.625, "grad_norm": 0.06857369827275735, "learning_rate": 6.89419795221843e-07, "loss": 0.0179, "num_tokens": 187332912.0, "reward": 0.140625, "reward_std": 0.1374513953924179, "rewards/accuracy_reward/mean": 0.140625, "rewards/accuracy_reward/std": 0.3479743003845215, "step": 203 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.2265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2034.0, "completions/mean_length": 1037.267578125, "completions/mean_terminated_length": 741.1944580078125, "completions/min_length": 18.0, "completions/min_terminated_length": 18.0, "epoch": 0.06964239993172314, "frac_reward_zero_std": 0.46875, "grad_norm": 0.09079845289212818, "learning_rate": 6.928327645051194e-07, "loss": 0.0355, "num_tokens": 187936265.0, "reward": 0.16015625, "reward_std": 0.18469981849193573, "rewards/accuracy_reward/mean": 0.16015625, "rewards/accuracy_reward/std": 0.3671095669269562, "step": 204 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.34375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2032.0, "completions/mean_length": 1330.427734375, "completions/mean_terminated_length": 954.5565795898438, "completions/min_length": 37.0, "completions/min_terminated_length": 37.0, "epoch": 0.06998378424511394, "frac_reward_zero_std": 0.75, "grad_norm": 0.056106399432478124, "learning_rate": 6.962457337883959e-07, "loss": 0.0009, "num_tokens": 188690836.0, "reward": 0.078125, "reward_std": 0.09891510009765625, "rewards/accuracy_reward/mean": 0.078125, "rewards/accuracy_reward/std": 0.26863065361976624, "step": 205 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.349609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2044.0, "completions/mean_length": 1353.337890625, "completions/mean_terminated_length": 979.9309692382812, "completions/min_length": 31.0, "completions/min_terminated_length": 31.0, "epoch": 0.07032516855850474, "frac_reward_zero_std": 0.5625, "grad_norm": 0.06861247736470905, "learning_rate": 6.996587030716723e-07, "loss": 0.0265, "num_tokens": 189468993.0, "reward": 0.154296875, "reward_std": 0.18038935959339142, "rewards/accuracy_reward/mean": 0.154296875, "rewards/accuracy_reward/std": 0.36158639192581177, "step": 206 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.326171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2014.0, "completions/mean_length": 1290.751953125, "completions/mean_terminated_length": 924.2000122070312, "completions/min_length": 26.0, "completions/min_terminated_length": 26.0, "epoch": 0.07066655287189554, "frac_reward_zero_std": 0.78125, "grad_norm": 0.05447422615165889, "learning_rate": 7.030716723549488e-07, "loss": 0.001, "num_tokens": 190205250.0, "reward": 0.095703125, "reward_std": 0.09519492089748383, "rewards/accuracy_reward/mean": 0.095703125, "rewards/accuracy_reward/std": 0.2944713830947876, "step": 207 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.33203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2037.0, "completions/mean_length": 1281.751953125, "completions/mean_terminated_length": 900.868408203125, "completions/min_length": 91.0, "completions/min_terminated_length": 91.0, "epoch": 0.07100793718528634, "frac_reward_zero_std": 0.625, "grad_norm": 0.06166585256041533, "learning_rate": 7.064846416382251e-07, "loss": 0.0131, "num_tokens": 190934291.0, "reward": 0.1640625, "reward_std": 0.1321541965007782, "rewards/accuracy_reward/mean": 0.1640625, "rewards/accuracy_reward/std": 0.37069445848464966, "step": 208 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2021.0, "completions/mean_length": 1163.673828125, "completions/mean_terminated_length": 843.8111572265625, "completions/min_length": 24.0, "completions/min_terminated_length": 24.0, "epoch": 0.07134932149867713, "frac_reward_zero_std": 0.625, "grad_norm": 0.06370224490183188, "learning_rate": 7.098976109215017e-07, "loss": 0.0071, "num_tokens": 191604732.0, "reward": 0.15625, "reward_std": 0.1473877727985382, "rewards/accuracy_reward/mean": 0.15625, "rewards/accuracy_reward/std": 0.36344730854034424, "step": 209 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.255859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2036.0, "completions/mean_length": 1182.1640625, "completions/mean_terminated_length": 884.4619140625, "completions/min_length": 113.0, "completions/min_terminated_length": 113.0, "epoch": 0.07169070581206793, "frac_reward_zero_std": 0.625, "grad_norm": 0.07316503857103573, "learning_rate": 7.133105802047781e-07, "loss": 0.0287, "num_tokens": 192279264.0, "reward": 0.078125, "reward_std": 0.13055649399757385, "rewards/accuracy_reward/mean": 0.08064515888690948, "rewards/accuracy_reward/std": 0.2725643217563629, "step": 210 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.275390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2007.0, "completions/mean_length": 1186.552734375, "completions/mean_terminated_length": 859.1563110351562, "completions/min_length": 16.0, "completions/min_terminated_length": 16.0, "epoch": 0.07203209012545873, "frac_reward_zero_std": 0.4375, "grad_norm": 0.0930444194257224, "learning_rate": 7.167235494880546e-07, "loss": -0.0049, "num_tokens": 192963819.0, "reward": 0.177734375, "reward_std": 0.21451082825660706, "rewards/accuracy_reward/mean": 0.177734375, "rewards/accuracy_reward/std": 0.3826628625392914, "step": 211 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.330078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2048.0, "completions/mean_length": 1345.55859375, "completions/mean_terminated_length": 999.457763671875, "completions/min_length": 78.0, "completions/min_terminated_length": 78.0, "epoch": 0.07237347443884953, "frac_reward_zero_std": 0.6875, "grad_norm": 0.06156584796150671, "learning_rate": 7.201365187713311e-07, "loss": 0.0035, "num_tokens": 193734761.0, "reward": 0.10546875, "reward_std": 0.11740703880786896, "rewards/accuracy_reward/mean": 0.10546875, "rewards/accuracy_reward/std": 0.3074568510055542, "step": 212 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.50390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2044.0, "completions/mean_length": 1544.287109375, "completions/mean_terminated_length": 1032.6417236328125, "completions/min_length": 26.0, "completions/min_terminated_length": 26.0, "epoch": 0.07271485875224033, "frac_reward_zero_std": 0.5625, "grad_norm": 0.06710965589022971, "learning_rate": 7.235494880546075e-07, "loss": 0.0127, "num_tokens": 194612476.0, "reward": 0.1640625, "reward_std": 0.1723763346672058, "rewards/accuracy_reward/mean": 0.1640625, "rewards/accuracy_reward/std": 0.37069445848464966, "step": 213 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.36328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1303.27734375, "completions/mean_terminated_length": 878.5091552734375, "completions/min_length": 16.0, "completions/min_terminated_length": 16.0, "epoch": 0.07305624306563113, "frac_reward_zero_std": 0.8125, "grad_norm": 0.04227260922021009, "learning_rate": 7.269624573378839e-07, "loss": 0.0083, "num_tokens": 195351818.0, "reward": 0.025390625, "reward_std": 0.06354551017284393, "rewards/accuracy_reward/mean": 0.026209676638245583, "rewards/accuracy_reward/std": 0.1599196493625641, "step": 214 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.392578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2045.0, "completions/mean_length": 1416.05859375, "completions/mean_terminated_length": 1007.6334228515625, "completions/min_length": 11.0, "completions/min_terminated_length": 11.0, "epoch": 0.07339762737902193, "frac_reward_zero_std": 0.78125, "grad_norm": 0.04471491194159816, "learning_rate": 7.303754266211603e-07, "loss": 0.0054, "num_tokens": 196156232.0, "reward": 0.08984375, "reward_std": 0.08601218461990356, "rewards/accuracy_reward/mean": 0.08984375, "rewards/accuracy_reward/std": 0.2862374484539032, "step": 215 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.369140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2040.0, "completions/mean_length": 1412.931640625, "completions/mean_terminated_length": 1041.328125, "completions/min_length": 18.0, "completions/min_terminated_length": 18.0, "epoch": 0.07373901169241273, "frac_reward_zero_std": 0.71875, "grad_norm": 0.05314749310020337, "learning_rate": 7.337883959044369e-07, "loss": -0.0014, "num_tokens": 196955029.0, "reward": 0.130859375, "reward_std": 0.1138748899102211, "rewards/accuracy_reward/mean": 0.130859375, "rewards/accuracy_reward/std": 0.33757632970809937, "step": 216 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.3671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2045.0, "completions/mean_length": 1371.814453125, "completions/mean_terminated_length": 979.4598999023438, "completions/min_length": 24.0, "completions/min_terminated_length": 24.0, "epoch": 0.07408039600580353, "frac_reward_zero_std": 0.625, "grad_norm": 0.06795948970036529, "learning_rate": 7.372013651877133e-07, "loss": 0.0203, "num_tokens": 197733670.0, "reward": 0.181640625, "reward_std": 0.14458514750003815, "rewards/accuracy_reward/mean": 0.181640625, "rewards/accuracy_reward/std": 0.38592514395713806, "step": 217 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2045.0, "completions/mean_length": 1361.314453125, "completions/mean_terminated_length": 976.1005859375, "completions/min_length": 19.0, "completions/min_terminated_length": 19.0, "epoch": 0.07442178031919433, "frac_reward_zero_std": 0.625, "grad_norm": 0.06342866930215545, "learning_rate": 7.406143344709898e-07, "loss": 0.0278, "num_tokens": 198505079.0, "reward": 0.119140625, "reward_std": 0.1513601392507553, "rewards/accuracy_reward/mean": 0.119140625, "rewards/accuracy_reward/std": 0.32427072525024414, "step": 218 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.38671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2044.0, "completions/mean_length": 1344.80859375, "completions/mean_terminated_length": 901.5191040039062, "completions/min_length": 23.0, "completions/min_terminated_length": 23.0, "epoch": 0.07476316463258513, "frac_reward_zero_std": 0.6875, "grad_norm": 0.06039354300396888, "learning_rate": 7.440273037542661e-07, "loss": 0.0124, "num_tokens": 199284229.0, "reward": 0.125, "reward_std": 0.13164834678173065, "rewards/accuracy_reward/mean": 0.125, "rewards/accuracy_reward/std": 0.3310423493385315, "step": 219 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.37109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2039.0, "completions/mean_length": 1403.40625, "completions/mean_terminated_length": 1023.055908203125, "completions/min_length": 21.0, "completions/min_terminated_length": 21.0, "epoch": 0.07510454894597593, "frac_reward_zero_std": 0.65625, "grad_norm": 0.05641142355994621, "learning_rate": 7.474402730375426e-07, "loss": 0.0158, "num_tokens": 200083925.0, "reward": 0.109375, "reward_std": 0.1270298957824707, "rewards/accuracy_reward/mean": 0.109375, "rewards/accuracy_reward/std": 0.31241437792778015, "step": 220 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.34375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2037.0, "completions/mean_length": 1405.111328125, "completions/mean_terminated_length": 1068.360107421875, "completions/min_length": 17.0, "completions/min_terminated_length": 17.0, "epoch": 0.07544593325936673, "frac_reward_zero_std": 0.65625, "grad_norm": 0.060386715579833, "learning_rate": 7.50853242320819e-07, "loss": 0.0072, "num_tokens": 200876062.0, "reward": 0.15234375, "reward_std": 0.12851551175117493, "rewards/accuracy_reward/mean": 0.15234375, "rewards/accuracy_reward/std": 0.35970520973205566, "step": 221 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.330078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2040.0, "completions/mean_length": 1339.912109375, "completions/mean_terminated_length": 991.0291748046875, "completions/min_length": 28.0, "completions/min_terminated_length": 28.0, "epoch": 0.07578731757275753, "frac_reward_zero_std": 0.71875, "grad_norm": 0.05638395971828098, "learning_rate": 7.542662116040955e-07, "loss": 0.0287, "num_tokens": 201640785.0, "reward": 0.14453125, "reward_std": 0.1186123713850975, "rewards/accuracy_reward/mean": 0.14453125, "rewards/accuracy_reward/std": 0.35197147727012634, "step": 222 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.322265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.0, "completions/mean_length": 1306.1484375, "completions/mean_terminated_length": 953.394775390625, "completions/min_length": 11.0, "completions/min_terminated_length": 11.0, "epoch": 0.07612870188614833, "frac_reward_zero_std": 0.59375, "grad_norm": 0.07496764447714771, "learning_rate": 7.57679180887372e-07, "loss": 0.0193, "num_tokens": 202388093.0, "reward": 0.234375, "reward_std": 0.16398611664772034, "rewards/accuracy_reward/mean": 0.234375, "rewards/accuracy_reward/std": 0.42402184009552, "step": 223 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2037.0, "completions/mean_length": 1326.798828125, "completions/mean_terminated_length": 974.5842895507812, "completions/min_length": 25.0, "completions/min_terminated_length": 25.0, "epoch": 0.07647008619953913, "frac_reward_zero_std": 0.75, "grad_norm": 0.050537073261015864, "learning_rate": 7.610921501706485e-07, "loss": -0.0044, "num_tokens": 203144518.0, "reward": 0.0859375, "reward_std": 0.08450747281312943, "rewards/accuracy_reward/mean": 0.0859375, "rewards/accuracy_reward/std": 0.28054583072662354, "step": 224 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2042.0, "completions/mean_length": 1483.244140625, "completions/mean_terminated_length": 1015.3035888671875, "completions/min_length": 121.0, "completions/min_terminated_length": 121.0, "epoch": 0.07681147051292993, "frac_reward_zero_std": 0.625, "grad_norm": 0.059106606224482036, "learning_rate": 7.645051194539249e-07, "loss": 0.0143, "num_tokens": 203984851.0, "reward": 0.12890625, "reward_std": 0.1426628828048706, "rewards/accuracy_reward/mean": 0.12890625, "rewards/accuracy_reward/std": 0.33542385697364807, "step": 225 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.34375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2041.0, "completions/mean_length": 1285.1796875, "completions/mean_terminated_length": 885.607177734375, "completions/min_length": 18.0, "completions/min_terminated_length": 18.0, "epoch": 0.07715285482632073, "frac_reward_zero_std": 0.625, "grad_norm": 0.082351681932432, "learning_rate": 7.679180887372013e-07, "loss": 0.0119, "num_tokens": 204726367.0, "reward": 0.126953125, "reward_std": 0.14634227752685547, "rewards/accuracy_reward/mean": 0.13104838132858276, "rewards/accuracy_reward/std": 0.3377939760684967, "step": 226 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.380859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2042.0, "completions/mean_length": 1400.884765625, "completions/mean_terminated_length": 1002.8170776367188, "completions/min_length": 25.0, "completions/min_terminated_length": 25.0, "epoch": 0.07749423913971153, "frac_reward_zero_std": 0.625, "grad_norm": 0.06143391851496004, "learning_rate": 7.713310580204778e-07, "loss": 0.0109, "num_tokens": 205522788.0, "reward": 0.115234375, "reward_std": 0.1425960510969162, "rewards/accuracy_reward/mean": 0.115234375, "rewards/accuracy_reward/std": 0.3196168541908264, "step": 227 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.384765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2039.0, "completions/mean_length": 1389.75390625, "completions/mean_terminated_length": 978.0889282226562, "completions/min_length": 34.0, "completions/min_terminated_length": 34.0, "epoch": 0.07783562345310233, "frac_reward_zero_std": 0.5625, "grad_norm": 0.06857189924028634, "learning_rate": 7.747440273037542e-07, "loss": -0.0069, "num_tokens": 206305542.0, "reward": 0.119140625, "reward_std": 0.17003285884857178, "rewards/accuracy_reward/mean": 0.119140625, "rewards/accuracy_reward/std": 0.32427072525024414, "step": 228 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2031.0, "completions/mean_length": 1446.267578125, "completions/mean_terminated_length": 1060.5416259765625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.07817700776649313, "frac_reward_zero_std": 0.59375, "grad_norm": 0.06060487939431986, "learning_rate": 7.781569965870307e-07, "loss": 0.026, "num_tokens": 207130655.0, "reward": 0.185546875, "reward_std": 0.1637648046016693, "rewards/accuracy_reward/mean": 0.19153225421905518, "rewards/accuracy_reward/std": 0.3939041793346405, "step": 229 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.3828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2024.0, "completions/mean_length": 1443.86328125, "completions/mean_terminated_length": 1069.1456298828125, "completions/min_length": 90.0, "completions/min_terminated_length": 90.0, "epoch": 0.07851839207988393, "frac_reward_zero_std": 0.75, "grad_norm": 0.04680176086940241, "learning_rate": 7.815699658703071e-07, "loss": 0.0128, "num_tokens": 207953817.0, "reward": 0.078125, "reward_std": 0.1085379421710968, "rewards/accuracy_reward/mean": 0.08064515888690948, "rewards/accuracy_reward/std": 0.2725643217563629, "step": 230 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.412109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2045.0, "completions/mean_length": 1362.62890625, "completions/mean_terminated_length": 882.18603515625, "completions/min_length": 10.0, "completions/min_terminated_length": 10.0, "epoch": 0.07885977639327472, "frac_reward_zero_std": 0.71875, "grad_norm": 0.05815743683902764, "learning_rate": 7.849829351535836e-07, "loss": 0.0162, "num_tokens": 208727835.0, "reward": 0.134765625, "reward_std": 0.11187098175287247, "rewards/accuracy_reward/mean": 0.134765625, "rewards/accuracy_reward/std": 0.3418070077896118, "step": 231 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.318359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2034.0, "completions/mean_length": 1386.005859375, "completions/mean_terminated_length": 1076.8223876953125, "completions/min_length": 55.0, "completions/min_terminated_length": 55.0, "epoch": 0.07920116070666552, "frac_reward_zero_std": 0.8125, "grad_norm": 0.04254597491119421, "learning_rate": 7.8839590443686e-07, "loss": 0.0134, "num_tokens": 209517294.0, "reward": 0.08203125, "reward_std": 0.077679343521595, "rewards/accuracy_reward/mean": 0.08203125, "rewards/accuracy_reward/std": 0.2746807038784027, "step": 232 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.400390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2044.0, "completions/mean_length": 1442.533203125, "completions/mean_terminated_length": 1038.231201171875, "completions/min_length": 159.0, "completions/min_terminated_length": 159.0, "epoch": 0.07954254502005632, "frac_reward_zero_std": 0.53125, "grad_norm": 0.06810755401398516, "learning_rate": 7.918088737201365e-07, "loss": -0.001, "num_tokens": 210331695.0, "reward": 0.19921875, "reward_std": 0.20368652045726776, "rewards/accuracy_reward/mean": 0.19921875, "rewards/accuracy_reward/std": 0.39980348944664, "step": 233 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.443359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2026.0, "completions/mean_length": 1470.01171875, "completions/mean_terminated_length": 1009.649169921875, "completions/min_length": 49.0, "completions/min_terminated_length": 49.0, "epoch": 0.07988392933344712, "frac_reward_zero_std": 0.84375, "grad_norm": 0.03783645855051803, "learning_rate": 7.952218430034129e-07, "loss": 0.0089, "num_tokens": 211165221.0, "reward": 0.029296875, "reward_std": 0.059079915285110474, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 234 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.404296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2044.0, "completions/mean_length": 1472.572265625, "completions/mean_terminated_length": 1082.0360107421875, "completions/min_length": 28.0, "completions/min_terminated_length": 28.0, "epoch": 0.08022531364683792, "frac_reward_zero_std": 0.59375, "grad_norm": 0.05560621980418969, "learning_rate": 7.986348122866893e-07, "loss": 0.0114, "num_tokens": 211991850.0, "reward": 0.142578125, "reward_std": 0.1544896960258484, "rewards/accuracy_reward/mean": 0.142578125, "rewards/accuracy_reward/std": 0.3499840497970581, "step": 235 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.34765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2044.0, "completions/mean_length": 1395.498046875, "completions/mean_terminated_length": 1047.757568359375, "completions/min_length": 28.0, "completions/min_terminated_length": 28.0, "epoch": 0.08056669796022872, "frac_reward_zero_std": 0.625, "grad_norm": 0.05202985810272377, "learning_rate": 8.020477815699659e-07, "loss": 0.0306, "num_tokens": 212780729.0, "reward": 0.16796875, "reward_std": 0.14055965840816498, "rewards/accuracy_reward/mean": 0.16796875, "rewards/accuracy_reward/std": 0.374204158782959, "step": 236 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.412109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1510.9765625, "completions/mean_terminated_length": 1134.52490234375, "completions/min_length": 66.0, "completions/min_terminated_length": 66.0, "epoch": 0.08090808227361952, "frac_reward_zero_std": 0.65625, "grad_norm": 0.05102068515338055, "learning_rate": 8.054607508532423e-07, "loss": 0.0064, "num_tokens": 213637437.0, "reward": 0.162109375, "reward_std": 0.13695251941680908, "rewards/accuracy_reward/mean": 0.162109375, "rewards/accuracy_reward/std": 0.3689115643501282, "step": 237 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.4453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2031.0, "completions/mean_length": 1426.62890625, "completions/mean_terminated_length": 927.7816772460938, "completions/min_length": 27.0, "completions/min_terminated_length": 27.0, "epoch": 0.08124946658701032, "frac_reward_zero_std": 0.8125, "grad_norm": 0.04009850102602469, "learning_rate": 8.088737201365188e-07, "loss": 0.007, "num_tokens": 214441935.0, "reward": 0.09375, "reward_std": 0.07582925260066986, "rewards/accuracy_reward/mean": 0.09375, "rewards/accuracy_reward/std": 0.29176566004753113, "step": 238 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2048.0, "completions/mean_length": 1509.158203125, "completions/mean_terminated_length": 1163.746826171875, "completions/min_length": 107.0, "completions/min_terminated_length": 107.0, "epoch": 0.08159085090040112, "frac_reward_zero_std": 0.5625, "grad_norm": 0.07606589914072429, "learning_rate": 8.122866894197952e-07, "loss": 0.0301, "num_tokens": 215295296.0, "reward": 0.1875, "reward_std": 0.19631412625312805, "rewards/accuracy_reward/mean": 0.1875, "rewards/accuracy_reward/std": 0.39069411158561707, "step": 239 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.431640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2043.0, "completions/mean_length": 1471.509765625, "completions/mean_terminated_length": 1033.6942138671875, "completions/min_length": 35.0, "completions/min_terminated_length": 35.0, "epoch": 0.08193223521379192, "frac_reward_zero_std": 0.625, "grad_norm": 0.05453874474922659, "learning_rate": 8.156996587030717e-07, "loss": 0.0143, "num_tokens": 216122181.0, "reward": 0.1484375, "reward_std": 0.14860516786575317, "rewards/accuracy_reward/mean": 0.15322580933570862, "rewards/accuracy_reward/std": 0.36056873202323914, "step": 240 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.3515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2023.0, "completions/mean_length": 1247.712890625, "completions/mean_terminated_length": 813.822265625, "completions/min_length": 37.0, "completions/min_terminated_length": 37.0, "epoch": 0.08227361952718272, "frac_reward_zero_std": 0.6875, "grad_norm": 0.06660764839208902, "learning_rate": 8.19112627986348e-07, "loss": -0.0137, "num_tokens": 216834370.0, "reward": 0.15625, "reward_std": 0.12367744743824005, "rewards/accuracy_reward/mean": 0.15625, "rewards/accuracy_reward/std": 0.36344730854034424, "step": 241 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.34375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2044.0, "completions/mean_length": 1350.517578125, "completions/mean_terminated_length": 985.169677734375, "completions/min_length": 28.0, "completions/min_terminated_length": 28.0, "epoch": 0.08261500384057352, "frac_reward_zero_std": 0.71875, "grad_norm": 0.06341202553911253, "learning_rate": 8.225255972696245e-07, "loss": 0.0159, "num_tokens": 217604331.0, "reward": 0.1171875, "reward_std": 0.10480518639087677, "rewards/accuracy_reward/mean": 0.1171875, "rewards/accuracy_reward/std": 0.32195815443992615, "step": 242 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.3828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2034.0, "completions/mean_length": 1428.5546875, "completions/mean_terminated_length": 1044.341796875, "completions/min_length": 157.0, "completions/min_terminated_length": 157.0, "epoch": 0.08295638815396432, "frac_reward_zero_std": 0.5625, "grad_norm": 0.06168287426030444, "learning_rate": 8.259385665529009e-07, "loss": 0.0228, "num_tokens": 218415223.0, "reward": 0.13671875, "reward_std": 0.1691007912158966, "rewards/accuracy_reward/mean": 0.13671875, "rewards/accuracy_reward/std": 0.3438861668109894, "step": 243 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.408203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1517.041015625, "completions/mean_terminated_length": 1150.802001953125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.08329777246735512, "frac_reward_zero_std": 0.59375, "grad_norm": 0.061142768295055536, "learning_rate": 8.293515358361775e-07, "loss": 0.0086, "num_tokens": 219270588.0, "reward": 0.12890625, "reward_std": 0.1570894569158554, "rewards/accuracy_reward/mean": 0.12890625, "rewards/accuracy_reward/std": 0.33542385697364807, "step": 244 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.3671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2034.0, "completions/mean_length": 1333.083984375, "completions/mean_terminated_length": 918.2561645507812, "completions/min_length": 8.0, "completions/min_terminated_length": 8.0, "epoch": 0.08363915678074592, "frac_reward_zero_std": 0.71875, "grad_norm": 0.05292730748114597, "learning_rate": 8.327645051194539e-07, "loss": 0.0052, "num_tokens": 220027383.0, "reward": 0.103515625, "reward_std": 0.10779774188995361, "rewards/accuracy_reward/mean": 0.103515625, "rewards/accuracy_reward/std": 0.30492907762527466, "step": 245 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.5078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2016.0, "completions/mean_length": 1571.427734375, "completions/mean_terminated_length": 1079.7261962890625, "completions/min_length": 31.0, "completions/min_terminated_length": 31.0, "epoch": 0.08398054109413672, "frac_reward_zero_std": 0.75, "grad_norm": 0.042786488185753105, "learning_rate": 8.361774744027303e-07, "loss": 0.0056, "num_tokens": 220911106.0, "reward": 0.078125, "reward_std": 0.09244653582572937, "rewards/accuracy_reward/mean": 0.078125, "rewards/accuracy_reward/std": 0.26863065361976624, "step": 246 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.392578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2045.0, "completions/mean_length": 1447.17578125, "completions/mean_terminated_length": 1058.8616943359375, "completions/min_length": 26.0, "completions/min_terminated_length": 26.0, "epoch": 0.08432192540752752, "frac_reward_zero_std": 0.46875, "grad_norm": 0.08304467300965157, "learning_rate": 8.395904436860067e-07, "loss": 0.0114, "num_tokens": 221733852.0, "reward": 0.17578125, "reward_std": 0.19873401522636414, "rewards/accuracy_reward/mean": 0.17578125, "rewards/accuracy_reward/std": 0.3810062110424042, "step": 247 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.478515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2034.0, "completions/mean_length": 1502.107421875, "completions/mean_terminated_length": 1001.1947631835938, "completions/min_length": 24.0, "completions/min_terminated_length": 24.0, "epoch": 0.08466330972091832, "frac_reward_zero_std": 0.71875, "grad_norm": 0.041970010222907216, "learning_rate": 8.430034129692832e-07, "loss": 0.0051, "num_tokens": 222584451.0, "reward": 0.12109375, "reward_std": 0.09122256934642792, "rewards/accuracy_reward/mean": 0.12109375, "rewards/accuracy_reward/std": 0.3265552520751953, "step": 248 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.4609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1499.66796875, "completions/mean_terminated_length": 1030.8043212890625, "completions/min_length": 64.0, "completions/min_terminated_length": 64.0, "epoch": 0.08500469403430912, "frac_reward_zero_std": 0.71875, "grad_norm": 0.04845680798453503, "learning_rate": 8.464163822525597e-07, "loss": 0.0139, "num_tokens": 223433161.0, "reward": 0.08984375, "reward_std": 0.10970106720924377, "rewards/accuracy_reward/mean": 0.0927419364452362, "rewards/accuracy_reward/std": 0.2903633117675781, "step": 249 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.373046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2045.0, "completions/mean_length": 1382.42578125, "completions/mean_terminated_length": 986.3987426757812, "completions/min_length": 33.0, "completions/min_terminated_length": 33.0, "epoch": 0.08534607834769992, "frac_reward_zero_std": 0.625, "grad_norm": 0.05890408312330982, "learning_rate": 8.498293515358362e-07, "loss": 0.0045, "num_tokens": 224213107.0, "reward": 0.138671875, "reward_std": 0.13934782147407532, "rewards/accuracy_reward/mean": 0.14791665971279144, "rewards/accuracy_reward/std": 0.35538771748542786, "step": 250 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.4453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1559.04296875, "completions/mean_terminated_length": 1166.5, "completions/min_length": 28.0, "completions/min_terminated_length": 28.0, "epoch": 0.08568746266109073, "frac_reward_zero_std": 0.6875, "grad_norm": 0.04489745088402795, "learning_rate": 8.532423208191127e-07, "loss": 0.0038, "num_tokens": 225086889.0, "reward": 0.13671875, "reward_std": 0.13031890988349915, "rewards/accuracy_reward/mean": 0.13671875, "rewards/accuracy_reward/std": 0.3438861668109894, "step": 251 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.4140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.0, "completions/mean_length": 1456.076171875, "completions/mean_terminated_length": 1037.9300537109375, "completions/min_length": 21.0, "completions/min_terminated_length": 21.0, "epoch": 0.08602884697448153, "frac_reward_zero_std": 0.75, "grad_norm": 0.040106339863035766, "learning_rate": 8.56655290102389e-07, "loss": -0.0013, "num_tokens": 225919072.0, "reward": 0.041015625, "reward_std": 0.09016455709934235, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 252 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.501953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2045.0, "completions/mean_length": 1567.36328125, "completions/mean_terminated_length": 1082.9569091796875, "completions/min_length": 11.0, "completions/min_terminated_length": 11.0, "epoch": 0.08637023128787233, "frac_reward_zero_std": 0.75, "grad_norm": 0.04487838386480456, "learning_rate": 8.600682593856655e-07, "loss": -0.0016, "num_tokens": 226810394.0, "reward": 0.080078125, "reward_std": 0.08642987906932831, "rewards/accuracy_reward/mean": 0.080078125, "rewards/accuracy_reward/std": 0.271679550409317, "step": 253 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.443359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2029.0, "completions/mean_length": 1479.53125, "completions/mean_terminated_length": 1026.7508544921875, "completions/min_length": 78.0, "completions/min_terminated_length": 78.0, "epoch": 0.08671161560126313, "frac_reward_zero_std": 0.5, "grad_norm": 0.061708670608455446, "learning_rate": 8.634812286689419e-07, "loss": 0.0066, "num_tokens": 227636954.0, "reward": 0.240234375, "reward_std": 0.1975778490304947, "rewards/accuracy_reward/mean": 0.240234375, "rewards/accuracy_reward/std": 0.4276435375213623, "step": 254 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.40234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2027.0, "completions/mean_length": 1446.666015625, "completions/mean_terminated_length": 1041.846435546875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.08705299991465393, "frac_reward_zero_std": 0.65625, "grad_norm": 0.059667980541240034, "learning_rate": 8.668941979522184e-07, "loss": 0.0097, "num_tokens": 228461391.0, "reward": 0.123046875, "reward_std": 0.14497441053390503, "rewards/accuracy_reward/mean": 0.1270161271095276, "rewards/accuracy_reward/std": 0.3333272337913513, "step": 255 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.447265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1510.134765625, "completions/mean_terminated_length": 1074.901123046875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.08739438422804473, "frac_reward_zero_std": 0.8125, "grad_norm": 0.03833520472893546, "learning_rate": 8.703071672354948e-07, "loss": 0.0091, "num_tokens": 229314084.0, "reward": 0.12890625, "reward_std": 0.08792105317115784, "rewards/accuracy_reward/mean": 0.12890625, "rewards/accuracy_reward/std": 0.33542385697364807, "step": 256 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.40625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2032.0, "completions/mean_length": 1529.36328125, "completions/mean_terminated_length": 1174.506591796875, "completions/min_length": 259.0, "completions/min_terminated_length": 259.0, "epoch": 0.08773576854143553, "frac_reward_zero_std": 0.5625, "grad_norm": 0.06010643865404402, "learning_rate": 8.737201365187713e-07, "loss": 0.0004, "num_tokens": 230172318.0, "reward": 0.193359375, "reward_std": 0.16956712305545807, "rewards/accuracy_reward/mean": 0.193359375, "rewards/accuracy_reward/std": 0.39531853795051575, "step": 257 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2044.0, "completions/mean_length": 1643.654296875, "completions/mean_terminated_length": 1263.814453125, "completions/min_length": 250.0, "completions/min_terminated_length": 250.0, "epoch": 0.08807715285482633, "frac_reward_zero_std": 0.75, "grad_norm": 0.03753888562273841, "learning_rate": 8.771331058020477e-07, "loss": 0.0119, "num_tokens": 231082845.0, "reward": 0.10546875, "reward_std": 0.11004265397787094, "rewards/accuracy_reward/mean": 0.10546875, "rewards/accuracy_reward/std": 0.3074568510055542, "step": 258 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.505859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2040.0, "completions/mean_length": 1573.89453125, "completions/mean_terminated_length": 1088.5455322265625, "completions/min_length": 29.0, "completions/min_terminated_length": 29.0, "epoch": 0.08841853716821713, "frac_reward_zero_std": 0.6875, "grad_norm": 0.04991002387235712, "learning_rate": 8.805460750853242e-07, "loss": 0.001, "num_tokens": 231963111.0, "reward": 0.115234375, "reward_std": 0.13447695970535278, "rewards/accuracy_reward/mean": 0.115234375, "rewards/accuracy_reward/std": 0.3196168541908264, "step": 259 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.482421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2039.0, "completions/mean_length": 1534.505859375, "completions/mean_terminated_length": 1055.890625, "completions/min_length": 28.0, "completions/min_terminated_length": 28.0, "epoch": 0.08875992148160793, "frac_reward_zero_std": 0.75, "grad_norm": 0.04476212872883397, "learning_rate": 8.839590443686007e-07, "loss": 0.0078, "num_tokens": 232828042.0, "reward": 0.03515625, "reward_std": 0.08493967354297638, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 260 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.50390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2038.0, "completions/mean_length": 1550.263671875, "completions/mean_terminated_length": 1044.68896484375, "completions/min_length": 26.0, "completions/min_terminated_length": 26.0, "epoch": 0.08910130579499873, "frac_reward_zero_std": 0.71875, "grad_norm": 0.05165846026565672, "learning_rate": 8.873720136518771e-07, "loss": -0.0062, "num_tokens": 233695969.0, "reward": 0.087890625, "reward_std": 0.10202304273843765, "rewards/accuracy_reward/mean": 0.087890625, "rewards/accuracy_reward/std": 0.2834126651287079, "step": 261 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.353515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2021.0, "completions/mean_length": 1383.859375, "completions/mean_terminated_length": 1020.6888427734375, "completions/min_length": 24.0, "completions/min_terminated_length": 24.0, "epoch": 0.08944269010838952, "frac_reward_zero_std": 0.53125, "grad_norm": 0.07046557420444258, "learning_rate": 8.907849829351535e-07, "loss": 0.039, "num_tokens": 234479433.0, "reward": 0.18359375, "reward_std": 0.18688803911209106, "rewards/accuracy_reward/mean": 0.18359375, "rewards/accuracy_reward/std": 0.3875311613082886, "step": 262 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.400390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.0, "completions/mean_length": 1436.2109375, "completions/mean_terminated_length": 1027.687255859375, "completions/min_length": 42.0, "completions/min_terminated_length": 42.0, "epoch": 0.08978407442178032, "frac_reward_zero_std": 0.53125, "grad_norm": 0.05845658105319272, "learning_rate": 8.941979522184299e-07, "loss": 0.0247, "num_tokens": 235288549.0, "reward": 0.18359375, "reward_std": 0.18581107258796692, "rewards/accuracy_reward/mean": 0.18359375, "rewards/accuracy_reward/std": 0.3875311613082886, "step": 263 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.482421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2043.0, "completions/mean_length": 1513.70703125, "completions/mean_terminated_length": 1015.7056884765625, "completions/min_length": 15.0, "completions/min_terminated_length": 15.0, "epoch": 0.09012545873517112, "frac_reward_zero_std": 0.75, "grad_norm": 0.05252347749924365, "learning_rate": 8.976109215017065e-07, "loss": -0.013, "num_tokens": 236147135.0, "reward": 0.08984375, "reward_std": 0.09319227933883667, "rewards/accuracy_reward/mean": 0.08984375, "rewards/accuracy_reward/std": 0.2862374484539032, "step": 264 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.560546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2045.0, "completions/mean_length": 1599.451171875, "completions/mean_terminated_length": 1027.30224609375, "completions/min_length": 24.0, "completions/min_terminated_length": 24.0, "epoch": 0.09046684304856192, "frac_reward_zero_std": 0.65625, "grad_norm": 0.056581265366370447, "learning_rate": 9.010238907849829e-07, "loss": 0.0214, "num_tokens": 237047766.0, "reward": 0.1171875, "reward_std": 0.12962549924850464, "rewards/accuracy_reward/mean": 0.1171875, "rewards/accuracy_reward/std": 0.32195815443992615, "step": 265 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.37890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1544.4765625, "completions/mean_terminated_length": 1237.2955322265625, "completions/min_length": 27.0, "completions/min_terminated_length": 27.0, "epoch": 0.09080822736195272, "frac_reward_zero_std": 0.5625, "grad_norm": 0.05766085684615005, "learning_rate": 9.044368600682594e-07, "loss": 0.0043, "num_tokens": 237913914.0, "reward": 0.15234375, "reward_std": 0.17622213065624237, "rewards/accuracy_reward/mean": 0.15234375, "rewards/accuracy_reward/std": 0.35970520973205566, "step": 266 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.40625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1478.552734375, "completions/mean_terminated_length": 1088.930908203125, "completions/min_length": 156.0, "completions/min_terminated_length": 156.0, "epoch": 0.09114961167534352, "frac_reward_zero_std": 0.59375, "grad_norm": 0.06155848409463368, "learning_rate": 9.078498293515358e-07, "loss": 0.0064, "num_tokens": 238749141.0, "reward": 0.216796875, "reward_std": 0.15439340472221375, "rewards/accuracy_reward/mean": 0.216796875, "rewards/accuracy_reward/std": 0.4124660789966583, "step": 267 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.509765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2048.0, "completions/mean_length": 1545.076171875, "completions/mean_terminated_length": 1022.1155395507812, "completions/min_length": 18.0, "completions/min_terminated_length": 18.0, "epoch": 0.09149099598873432, "frac_reward_zero_std": 0.59375, "grad_norm": 0.060196660541649086, "learning_rate": 9.112627986348122e-07, "loss": 0.0132, "num_tokens": 239616972.0, "reward": 0.138671875, "reward_std": 0.12972497940063477, "rewards/accuracy_reward/mean": 0.138671875, "rewards/accuracy_reward/std": 0.34594178199768066, "step": 268 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.5234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2042.0, "completions/mean_length": 1612.6328125, "completions/mean_terminated_length": 1134.4425048828125, "completions/min_length": 235.0, "completions/min_terminated_length": 235.0, "epoch": 0.09183238030212512, "frac_reward_zero_std": 0.5625, "grad_norm": 0.05984180948321153, "learning_rate": 9.146757679180886e-07, "loss": 0.03, "num_tokens": 240515408.0, "reward": 0.1640625, "reward_std": 0.17900297045707703, "rewards/accuracy_reward/mean": 0.1640625, "rewards/accuracy_reward/std": 0.37069445848464966, "step": 269 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.48046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2043.0, "completions/mean_length": 1497.748046875, "completions/mean_terminated_length": 988.8684692382812, "completions/min_length": 44.0, "completions/min_terminated_length": 44.0, "epoch": 0.09217376461551592, "frac_reward_zero_std": 0.625, "grad_norm": 0.056143753697885226, "learning_rate": 9.180887372013651e-07, "loss": 0.0178, "num_tokens": 241353615.0, "reward": 0.177734375, "reward_std": 0.14117157459259033, "rewards/accuracy_reward/mean": 0.177734375, "rewards/accuracy_reward/std": 0.3826628625392914, "step": 270 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.45703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2036.0, "completions/mean_length": 1521.458984375, "completions/mean_terminated_length": 1078.25537109375, "completions/min_length": 8.0, "completions/min_terminated_length": 8.0, "epoch": 0.09251514892890672, "frac_reward_zero_std": 0.65625, "grad_norm": 0.05255470997158037, "learning_rate": 9.215017064846417e-07, "loss": 0.0166, "num_tokens": 242206522.0, "reward": 0.177734375, "reward_std": 0.13407659530639648, "rewards/accuracy_reward/mean": 0.177734375, "rewards/accuracy_reward/std": 0.3826628625392914, "step": 271 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.43359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2044.0, "completions/mean_length": 1488.119140625, "completions/mean_terminated_length": 1059.5206298828125, "completions/min_length": 34.0, "completions/min_terminated_length": 34.0, "epoch": 0.09285653324229752, "frac_reward_zero_std": 0.6875, "grad_norm": 0.05285557067370238, "learning_rate": 9.249146757679181e-07, "loss": 0.0075, "num_tokens": 243042535.0, "reward": 0.14453125, "reward_std": 0.1233629435300827, "rewards/accuracy_reward/mean": 0.14453125, "rewards/accuracy_reward/std": 0.35197147727012634, "step": 272 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.51171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2034.0, "completions/mean_length": 1511.04296875, "completions/mean_terminated_length": 948.3120727539062, "completions/min_length": 30.0, "completions/min_terminated_length": 30.0, "epoch": 0.09319791755568832, "frac_reward_zero_std": 0.59375, "grad_norm": 0.05525082051472725, "learning_rate": 9.283276450511945e-07, "loss": 0.0243, "num_tokens": 243894253.0, "reward": 0.189453125, "reward_std": 0.1722542941570282, "rewards/accuracy_reward/mean": 0.189453125, "rewards/accuracy_reward/std": 0.3922513723373413, "step": 273 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.48046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2043.0, "completions/mean_length": 1517.673828125, "completions/mean_terminated_length": 1027.2218017578125, "completions/min_length": 148.0, "completions/min_terminated_length": 148.0, "epoch": 0.09353930186907912, "frac_reward_zero_std": 0.75, "grad_norm": 0.04270737137810146, "learning_rate": 9.317406143344709e-07, "loss": 0.0083, "num_tokens": 244742678.0, "reward": 0.080078125, "reward_std": 0.10896982997655869, "rewards/accuracy_reward/mean": 0.08266129344701767, "rewards/accuracy_reward/std": 0.2756476104259491, "step": 274 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.41796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2030.0, "completions/mean_length": 1423.15625, "completions/mean_terminated_length": 974.4429321289062, "completions/min_length": 27.0, "completions/min_terminated_length": 27.0, "epoch": 0.09388068618246992, "frac_reward_zero_std": 0.6875, "grad_norm": 0.04934473310328268, "learning_rate": 9.351535836177474e-07, "loss": 0.0129, "num_tokens": 245554390.0, "reward": 0.126953125, "reward_std": 0.1299196481704712, "rewards/accuracy_reward/mean": 0.13104838132858276, "rewards/accuracy_reward/std": 0.3377939760684967, "step": 275 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.419921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.0, "completions/mean_length": 1535.0546875, "completions/mean_terminated_length": 1163.7305908203125, "completions/min_length": 50.0, "completions/min_terminated_length": 50.0, "epoch": 0.09422207049586072, "frac_reward_zero_std": 0.59375, "grad_norm": 0.05899691461891417, "learning_rate": 9.385665529010238e-07, "loss": 0.0258, "num_tokens": 246412674.0, "reward": 0.255859375, "reward_std": 0.1746368557214737, "rewards/accuracy_reward/mean": 0.255859375, "rewards/accuracy_reward/std": 0.43676990270614624, "step": 276 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.466796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2044.0, "completions/mean_length": 1492.34375, "completions/mean_terminated_length": 1005.89013671875, "completions/min_length": 36.0, "completions/min_terminated_length": 36.0, "epoch": 0.09456345480925152, "frac_reward_zero_std": 0.6875, "grad_norm": 0.048357678495694646, "learning_rate": 9.419795221843004e-07, "loss": 0.0141, "num_tokens": 247251074.0, "reward": 0.134765625, "reward_std": 0.12556564807891846, "rewards/accuracy_reward/mean": 0.1391129046678543, "rewards/accuracy_reward/std": 0.34641367197036743, "step": 277 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.544921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2043.0, "completions/mean_length": 1628.275390625, "completions/mean_terminated_length": 1125.6866455078125, "completions/min_length": 29.0, "completions/min_terminated_length": 29.0, "epoch": 0.09490483912264232, "frac_reward_zero_std": 0.625, "grad_norm": 0.0517344088762816, "learning_rate": 9.453924914675768e-07, "loss": 0.0079, "num_tokens": 248160703.0, "reward": 0.162109375, "reward_std": 0.15084770321846008, "rewards/accuracy_reward/mean": 0.162109375, "rewards/accuracy_reward/std": 0.3689115643501282, "step": 278 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.587890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2044.0, "completions/mean_length": 1684.41796875, "completions/mean_terminated_length": 1165.9573974609375, "completions/min_length": 28.0, "completions/min_terminated_length": 28.0, "epoch": 0.09524622343603312, "frac_reward_zero_std": 0.71875, "grad_norm": 0.04413058747712155, "learning_rate": 9.488054607508532e-07, "loss": 0.0103, "num_tokens": 249110565.0, "reward": 0.111328125, "reward_std": 0.11575192213058472, "rewards/accuracy_reward/mean": 0.111328125, "rewards/accuracy_reward/std": 0.31484565138816833, "step": 279 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.48046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.0, "completions/mean_length": 1564.01953125, "completions/mean_terminated_length": 1116.4285888671875, "completions/min_length": 193.0, "completions/min_terminated_length": 193.0, "epoch": 0.09558760774942392, "frac_reward_zero_std": 0.75, "grad_norm": 0.04799554240173732, "learning_rate": 9.522184300341296e-07, "loss": 0.0057, "num_tokens": 249980143.0, "reward": 0.08203125, "reward_std": 0.0969250351190567, "rewards/accuracy_reward/mean": 0.08203125, "rewards/accuracy_reward/std": 0.2746807038784027, "step": 280 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.541015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1567.552734375, "completions/mean_terminated_length": 1001.23828125, "completions/min_length": 28.0, "completions/min_terminated_length": 28.0, "epoch": 0.09592899206281472, "frac_reward_zero_std": 0.65625, "grad_norm": 0.04377724093945442, "learning_rate": 9.556313993174062e-07, "loss": 0.0043, "num_tokens": 250857450.0, "reward": 0.13671875, "reward_std": 0.1284025013446808, "rewards/accuracy_reward/mean": 0.13671875, "rewards/accuracy_reward/std": 0.3438861668109894, "step": 281 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.470703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2039.0, "completions/mean_length": 1486.837890625, "completions/mean_terminated_length": 987.7970581054688, "completions/min_length": 28.0, "completions/min_terminated_length": 28.0, "epoch": 0.09627037637620552, "frac_reward_zero_std": 0.75, "grad_norm": 0.039757610300643106, "learning_rate": 9.590443686006826e-07, "loss": 0.0124, "num_tokens": 251689239.0, "reward": 0.158203125, "reward_std": 0.10262156277894974, "rewards/accuracy_reward/mean": 0.158203125, "rewards/accuracy_reward/std": 0.36528825759887695, "step": 282 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.560546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2041.0, "completions/mean_length": 1668.14453125, "completions/mean_terminated_length": 1183.6177978515625, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.09661176068959632, "frac_reward_zero_std": 0.75, "grad_norm": 0.03951743461345432, "learning_rate": 9.62457337883959e-07, "loss": 0.0101, "num_tokens": 252631617.0, "reward": 0.080078125, "reward_std": 0.1160123199224472, "rewards/accuracy_reward/mean": 0.080078125, "rewards/accuracy_reward/std": 0.271679550409317, "step": 283 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.521484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2034.0, "completions/mean_length": 1545.779296875, "completions/mean_terminated_length": 998.461181640625, "completions/min_length": 14.0, "completions/min_terminated_length": 14.0, "epoch": 0.09695314500298712, "frac_reward_zero_std": 0.65625, "grad_norm": 0.05704696957724095, "learning_rate": 9.658703071672355e-07, "loss": 0.0288, "num_tokens": 253491488.0, "reward": 0.1640625, "reward_std": 0.14701546728610992, "rewards/accuracy_reward/mean": 0.1640625, "rewards/accuracy_reward/std": 0.37069445848464966, "step": 284 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.587890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2032.0, "completions/mean_length": 1680.263671875, "completions/mean_terminated_length": 1155.6729736328125, "completions/min_length": 83.0, "completions/min_terminated_length": 83.0, "epoch": 0.09729452931637791, "frac_reward_zero_std": 0.5625, "grad_norm": 0.0690785121234627, "learning_rate": 9.69283276450512e-07, "loss": 0.022, "num_tokens": 254435175.0, "reward": 0.177734375, "reward_std": 0.19123506546020508, "rewards/accuracy_reward/mean": 0.177734375, "rewards/accuracy_reward/std": 0.3826628625392914, "step": 285 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.564453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2029.0, "completions/mean_length": 1595.998046875, "completions/mean_terminated_length": 1010.2197875976562, "completions/min_length": 28.0, "completions/min_terminated_length": 28.0, "epoch": 0.09763591362976871, "frac_reward_zero_std": 0.6875, "grad_norm": 0.04632495698952849, "learning_rate": 9.726962457337883e-07, "loss": 0.0206, "num_tokens": 255331670.0, "reward": 0.1640625, "reward_std": 0.12637948989868164, "rewards/accuracy_reward/mean": 0.1640625, "rewards/accuracy_reward/std": 0.37069445848464966, "step": 286 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.52734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2021.0, "completions/mean_length": 1582.466796875, "completions/mean_terminated_length": 1063.0701904296875, "completions/min_length": 105.0, "completions/min_terminated_length": 105.0, "epoch": 0.09797729794315951, "frac_reward_zero_std": 0.71875, "grad_norm": 0.04924463534046396, "learning_rate": 9.761092150170647e-07, "loss": 0.015, "num_tokens": 256228805.0, "reward": 0.078125, "reward_std": 0.10144393146038055, "rewards/accuracy_reward/mean": 0.078125, "rewards/accuracy_reward/std": 0.26863065361976624, "step": 287 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.521484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.0, "completions/mean_length": 1610.263671875, "completions/mean_terminated_length": 1133.2203369140625, "completions/min_length": 89.0, "completions/min_terminated_length": 89.0, "epoch": 0.09831868225655031, "frac_reward_zero_std": 0.71875, "grad_norm": 0.040754809714969024, "learning_rate": 9.795221843003413e-07, "loss": 0.0043, "num_tokens": 257134572.0, "reward": 0.083984375, "reward_std": 0.1062953919172287, "rewards/accuracy_reward/mean": 0.083984375, "rewards/accuracy_reward/std": 0.2776356339454651, "step": 288 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.626953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2034.0, "completions/mean_length": 1694.11328125, "completions/mean_terminated_length": 1099.361328125, "completions/min_length": 25.0, "completions/min_terminated_length": 25.0, "epoch": 0.09866006656994111, "frac_reward_zero_std": 0.625, "grad_norm": 0.04639383445604393, "learning_rate": 9.829351535836176e-07, "loss": 0.0024, "num_tokens": 258079670.0, "reward": 0.142578125, "reward_std": 0.150249183177948, "rewards/accuracy_reward/mean": 0.142578125, "rewards/accuracy_reward/std": 0.3499840497970581, "step": 289 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.521484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2043.0, "completions/mean_length": 1573.62890625, "completions/mean_terminated_length": 1056.6611328125, "completions/min_length": 20.0, "completions/min_terminated_length": 20.0, "epoch": 0.09900145088333191, "frac_reward_zero_std": 0.75, "grad_norm": 0.04219558774292927, "learning_rate": 9.863481228668942e-07, "loss": -0.0027, "num_tokens": 258961128.0, "reward": 0.119140625, "reward_std": 0.08941882103681564, "rewards/accuracy_reward/mean": 0.119140625, "rewards/accuracy_reward/std": 0.32427072525024414, "step": 290 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2044.0, "completions/mean_length": 1706.4921875, "completions/mean_terminated_length": 1238.5, "completions/min_length": 285.0, "completions/min_terminated_length": 285.0, "epoch": 0.09934283519672271, "frac_reward_zero_std": 0.75, "grad_norm": 0.040709983969185425, "learning_rate": 9.897610921501706e-07, "loss": 0.0203, "num_tokens": 259912292.0, "reward": 0.16015625, "reward_std": 0.09666463732719421, "rewards/accuracy_reward/mean": 0.16015625, "rewards/accuracy_reward/std": 0.3671095669269562, "step": 291 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.57421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1612.537109375, "completions/mean_terminated_length": 1025.261474609375, "completions/min_length": 25.0, "completions/min_terminated_length": 25.0, "epoch": 0.09968421951011351, "frac_reward_zero_std": 0.71875, "grad_norm": 0.04276884810934766, "learning_rate": 9.93174061433447e-07, "loss": 0.025, "num_tokens": 260813431.0, "reward": 0.1015625, "reward_std": 0.10107709467411041, "rewards/accuracy_reward/mean": 0.1015625, "rewards/accuracy_reward/std": 0.30236753821372986, "step": 292 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.52734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2043.0, "completions/mean_length": 1566.828125, "completions/mean_terminated_length": 1029.9833984375, "completions/min_length": 241.0, "completions/min_terminated_length": 241.0, "epoch": 0.10002560382350431, "frac_reward_zero_std": 0.5625, "grad_norm": 0.0550766362489316, "learning_rate": 9.965870307167234e-07, "loss": 0.0223, "num_tokens": 261691679.0, "reward": 0.166015625, "reward_std": 0.16999930143356323, "rewards/accuracy_reward/mean": 0.166015625, "rewards/accuracy_reward/std": 0.3724585771560669, "step": 293 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.5078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2039.0, "completions/mean_length": 1609.408203125, "completions/mean_terminated_length": 1156.8929443359375, "completions/min_length": 178.0, "completions/min_terminated_length": 178.0, "epoch": 0.10036698813689511, "frac_reward_zero_std": 0.59375, "grad_norm": 0.06305400424151338, "learning_rate": 1e-06, "loss": 0.0196, "num_tokens": 262592992.0, "reward": 0.109375, "reward_std": 0.14442548155784607, "rewards/accuracy_reward/mean": 0.109375, "rewards/accuracy_reward/std": 0.31241437792778015, "step": 294 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.59375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1683.541015625, "completions/mean_terminated_length": 1150.8702392578125, "completions/min_length": 210.0, "completions/min_terminated_length": 210.0, "epoch": 0.10070837245028591, "frac_reward_zero_std": 0.78125, "grad_norm": 0.03685004347579364, "learning_rate": 9.99999680653653e-07, "loss": 0.0112, "num_tokens": 263528741.0, "reward": 0.103515625, "reward_std": 0.10834437608718872, "rewards/accuracy_reward/mean": 0.103515625, "rewards/accuracy_reward/std": 0.30492907762527466, "step": 295 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.470703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2035.0, "completions/mean_length": 1464.427734375, "completions/mean_terminated_length": 945.4575805664062, "completions/min_length": 16.0, "completions/min_terminated_length": 16.0, "epoch": 0.10104975676367671, "frac_reward_zero_std": 0.53125, "grad_norm": 0.06555009323554643, "learning_rate": 9.999987226150655e-07, "loss": 0.0293, "num_tokens": 264350288.0, "reward": 0.18359375, "reward_std": 0.18672311305999756, "rewards/accuracy_reward/mean": 0.18359375, "rewards/accuracy_reward/std": 0.3875311613082886, "step": 296 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.666015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1743.46875, "completions/mean_terminated_length": 1136.1871337890625, "completions/min_length": 23.0, "completions/min_terminated_length": 23.0, "epoch": 0.10139114107706751, "frac_reward_zero_std": 0.625, "grad_norm": 0.05346895227262486, "learning_rate": 9.99997125885597e-07, "loss": 0.0148, "num_tokens": 265320928.0, "reward": 0.125, "reward_std": 0.1546129435300827, "rewards/accuracy_reward/mean": 0.125, "rewards/accuracy_reward/std": 0.3310423493385315, "step": 297 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.615234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2042.0, "completions/mean_length": 1677.2578125, "completions/mean_terminated_length": 1084.4466552734375, "completions/min_length": 51.0, "completions/min_terminated_length": 51.0, "epoch": 0.10173252539045831, "frac_reward_zero_std": 0.8125, "grad_norm": 0.03189958739910039, "learning_rate": 9.99994890467514e-07, "loss": 0.0066, "num_tokens": 266253252.0, "reward": 0.107421875, "reward_std": 0.07014618813991547, "rewards/accuracy_reward/mean": 0.107421875, "rewards/accuracy_reward/std": 0.30995169281959534, "step": 298 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.607421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2035.0, "completions/mean_length": 1679.44140625, "completions/mean_terminated_length": 1109.18408203125, "completions/min_length": 171.0, "completions/min_terminated_length": 171.0, "epoch": 0.10207390970384911, "frac_reward_zero_std": 0.625, "grad_norm": 0.06666519777262402, "learning_rate": 9.999920163639891e-07, "loss": 0.0149, "num_tokens": 267193526.0, "reward": 0.08984375, "reward_std": 0.12279174476861954, "rewards/accuracy_reward/mean": 0.08984375, "rewards/accuracy_reward/std": 0.2862374484539032, "step": 299 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.603515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.0, "completions/mean_length": 1647.5546875, "completions/mean_terminated_length": 1038.0098876953125, "completions/min_length": 16.0, "completions/min_terminated_length": 16.0, "epoch": 0.10241529401723991, "frac_reward_zero_std": 0.75, "grad_norm": 0.044562918221771934, "learning_rate": 9.999885035791019e-07, "loss": 0.0195, "num_tokens": 268118130.0, "reward": 0.08203125, "reward_std": 0.09946031868457794, "rewards/accuracy_reward/mean": 0.08203125, "rewards/accuracy_reward/std": 0.2746807038784027, "step": 300 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.572265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1697.453125, "completions/mean_terminated_length": 1228.45654296875, "completions/min_length": 200.0, "completions/min_terminated_length": 200.0, "epoch": 0.10275667833063071, "frac_reward_zero_std": 0.625, "grad_norm": 0.05308401775817464, "learning_rate": 9.999843521178375e-07, "loss": 0.0152, "num_tokens": 269061674.0, "reward": 0.171875, "reward_std": 0.1400686800479889, "rewards/accuracy_reward/mean": 0.171875, "rewards/accuracy_reward/std": 0.3776407241821289, "step": 301 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.65625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2027.0, "completions/mean_length": 1699.3046875, "completions/mean_terminated_length": 1033.6136474609375, "completions/min_length": 160.0, "completions/min_terminated_length": 160.0, "epoch": 0.10309806264402151, "frac_reward_zero_std": 0.6875, "grad_norm": 0.04676814457056771, "learning_rate": 9.99979561986089e-07, "loss": 0.0067, "num_tokens": 270009606.0, "reward": 0.12890625, "reward_std": 0.10876302421092987, "rewards/accuracy_reward/mean": 0.12890625, "rewards/accuracy_reward/std": 0.33542385697364807, "step": 302 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.68359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.0, "completions/mean_length": 1779.71875, "completions/mean_terminated_length": 1200.4197998046875, "completions/min_length": 71.0, "completions/min_terminated_length": 71.0, "epoch": 0.1034394469574123, "frac_reward_zero_std": 0.71875, "grad_norm": 0.044440036096316506, "learning_rate": 9.999741331906542e-07, "loss": 0.013, "num_tokens": 271005766.0, "reward": 0.1171875, "reward_std": 0.11595635861158371, "rewards/accuracy_reward/mean": 0.1171875, "rewards/accuracy_reward/std": 0.32195815443992615, "step": 303 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.66015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2045.0, "completions/mean_length": 1739.099609375, "completions/mean_terminated_length": 1139.0517578125, "completions/min_length": 60.0, "completions/min_terminated_length": 60.0, "epoch": 0.1037808312708031, "frac_reward_zero_std": 0.75, "grad_norm": 0.050629170052133174, "learning_rate": 9.99968065739239e-07, "loss": 0.0179, "num_tokens": 271972297.0, "reward": 0.123046875, "reward_std": 0.08224457502365112, "rewards/accuracy_reward/mean": 0.123046875, "rewards/accuracy_reward/std": 0.32881227135658264, "step": 304 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.685546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2015.0, "completions/mean_length": 1762.671875, "completions/mean_terminated_length": 1140.62109375, "completions/min_length": 172.0, "completions/min_terminated_length": 172.0, "epoch": 0.1041222155841939, "frac_reward_zero_std": 0.6875, "grad_norm": 0.047535662566425926, "learning_rate": 9.999613596404544e-07, "loss": 0.0178, "num_tokens": 272948241.0, "reward": 0.146484375, "reward_std": 0.11751356720924377, "rewards/accuracy_reward/mean": 0.146484375, "rewards/accuracy_reward/std": 0.35393697023391724, "step": 305 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1655.11328125, "completions/mean_terminated_length": 878.8488159179688, "completions/min_length": 145.0, "completions/min_terminated_length": 145.0, "epoch": 0.1044635998975847, "frac_reward_zero_std": 0.59375, "grad_norm": 0.058714538928285265, "learning_rate": 9.999540149038193e-07, "loss": 0.0141, "num_tokens": 273879083.0, "reward": 0.171875, "reward_std": 0.14168405532836914, "rewards/accuracy_reward/mean": 0.171875, "rewards/accuracy_reward/std": 0.3776407241821289, "step": 306 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.537109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2034.0, "completions/mean_length": 1527.306640625, "completions/mean_terminated_length": 923.1265258789062, "completions/min_length": 140.0, "completions/min_terminated_length": 140.0, "epoch": 0.1048049842109755, "frac_reward_zero_std": 0.71875, "grad_norm": 0.05408849262602303, "learning_rate": 9.999460315397577e-07, "loss": 0.0174, "num_tokens": 274733336.0, "reward": 0.08984375, "reward_std": 0.10480421781539917, "rewards/accuracy_reward/mean": 0.09583333134651184, "rewards/accuracy_reward/std": 0.29466965794563293, "step": 307 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.65625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2044.0, "completions/mean_length": 1725.140625, "completions/mean_terminated_length": 1108.772705078125, "completions/min_length": 30.0, "completions/min_terminated_length": 30.0, "epoch": 0.1051463685243663, "frac_reward_zero_std": 0.6875, "grad_norm": 0.04431010932684938, "learning_rate": 9.999374095596004e-07, "loss": 0.0085, "num_tokens": 275696576.0, "reward": 0.12109375, "reward_std": 0.1114048957824707, "rewards/accuracy_reward/mean": 0.12109375, "rewards/accuracy_reward/std": 0.3265552520751953, "step": 308 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.517578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2001.0, "completions/mean_length": 1502.541015625, "completions/mean_terminated_length": 917.33203125, "completions/min_length": 35.0, "completions/min_terminated_length": 35.0, "epoch": 0.1054877528377571, "frac_reward_zero_std": 0.6875, "grad_norm": 0.051023601659573634, "learning_rate": 9.99928148975585e-07, "loss": 0.0019, "num_tokens": 276542629.0, "reward": 0.13671875, "reward_std": 0.12973949313163757, "rewards/accuracy_reward/mean": 0.13671875, "rewards/accuracy_reward/std": 0.3438861668109894, "step": 309 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.517578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2041.0, "completions/mean_length": 1568.7578125, "completions/mean_terminated_length": 1054.591064453125, "completions/min_length": 124.0, "completions/min_terminated_length": 124.0, "epoch": 0.1058291371511479, "frac_reward_zero_std": 0.6875, "grad_norm": 0.05142377313796195, "learning_rate": 9.99918249800855e-07, "loss": 0.0131, "num_tokens": 277418153.0, "reward": 0.166015625, "reward_std": 0.11503897607326508, "rewards/accuracy_reward/mean": 0.166015625, "rewards/accuracy_reward/std": 0.3724585771560669, "step": 310 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.61328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1679.529296875, "completions/mean_terminated_length": 1095.1868896484375, "completions/min_length": 26.0, "completions/min_terminated_length": 26.0, "epoch": 0.1061705214645387, "frac_reward_zero_std": 0.78125, "grad_norm": 0.0398715324136502, "learning_rate": 9.999077120494608e-07, "loss": 0.002, "num_tokens": 278357240.0, "reward": 0.06640625, "reward_std": 0.06904878467321396, "rewards/accuracy_reward/mean": 0.06640625, "rewards/accuracy_reward/std": 0.2492343932390213, "step": 311 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.66015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2042.0, "completions/mean_length": 1679.583984375, "completions/mean_terminated_length": 963.92529296875, "completions/min_length": 45.0, "completions/min_terminated_length": 45.0, "epoch": 0.1065119057779295, "frac_reward_zero_std": 0.71875, "grad_norm": 0.04746502445588205, "learning_rate": 9.998965357363583e-07, "loss": 0.0113, "num_tokens": 279291731.0, "reward": 0.12109375, "reward_std": 0.116529181599617, "rewards/accuracy_reward/mean": 0.12109375, "rewards/accuracy_reward/std": 0.3265552520751953, "step": 312 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.63671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2037.0, "completions/mean_length": 1699.498046875, "completions/mean_terminated_length": 1088.6827392578125, "completions/min_length": 126.0, "completions/min_terminated_length": 126.0, "epoch": 0.1068532900913203, "frac_reward_zero_std": 0.625, "grad_norm": 0.05746511849257925, "learning_rate": 9.998847208774107e-07, "loss": 0.0223, "num_tokens": 280234018.0, "reward": 0.15234375, "reward_std": 0.1403878629207611, "rewards/accuracy_reward/mean": 0.15234375, "rewards/accuracy_reward/std": 0.35970520973205566, "step": 313 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.57421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2023.0, "completions/mean_length": 1676.560546875, "completions/mean_terminated_length": 1175.62841796875, "completions/min_length": 30.0, "completions/min_terminated_length": 30.0, "epoch": 0.1071946744047111, "frac_reward_zero_std": 0.6875, "grad_norm": 0.05129375878392557, "learning_rate": 9.998722674893869e-07, "loss": 0.0165, "num_tokens": 281171217.0, "reward": 0.14453125, "reward_std": 0.13770455121994019, "rewards/accuracy_reward/mean": 0.14453125, "rewards/accuracy_reward/std": 0.35197147727012634, "step": 314 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.646484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2023.0, "completions/mean_length": 1711.6640625, "completions/mean_terminated_length": 1096.91162109375, "completions/min_length": 50.0, "completions/min_terminated_length": 50.0, "epoch": 0.1075360587181019, "frac_reward_zero_std": 0.71875, "grad_norm": 0.04940003485514998, "learning_rate": 9.99859175589962e-07, "loss": 0.0227, "num_tokens": 282123077.0, "reward": 0.10546875, "reward_std": 0.11658906936645508, "rewards/accuracy_reward/mean": 0.1088709682226181, "rewards/accuracy_reward/std": 0.31179171800613403, "step": 315 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.697265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1796.013671875, "completions/mean_terminated_length": 1215.6322021484375, "completions/min_length": 49.0, "completions/min_terminated_length": 49.0, "epoch": 0.1078774430314927, "frac_reward_zero_std": 0.84375, "grad_norm": 0.030016141859968596, "learning_rate": 9.998454451977178e-07, "loss": -0.0008, "num_tokens": 283118684.0, "reward": 0.09765625, "reward_std": 0.06711846590042114, "rewards/accuracy_reward/mean": 0.09765625, "rewards/accuracy_reward/std": 0.29713961482048035, "step": 316 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.685546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2041.0, "completions/mean_length": 1741.78125, "completions/mean_terminated_length": 1074.1864013671875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1082188273448835, "frac_reward_zero_std": 0.65625, "grad_norm": 0.05766121026930601, "learning_rate": 9.99831076332142e-07, "loss": 0.0159, "num_tokens": 284085836.0, "reward": 0.142578125, "reward_std": 0.14102670550346375, "rewards/accuracy_reward/mean": 0.142578125, "rewards/accuracy_reward/std": 0.3499840497970581, "step": 317 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.5625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2035.0, "completions/mean_length": 1569.150390625, "completions/mean_terminated_length": 953.4866333007812, "completions/min_length": 114.0, "completions/min_terminated_length": 114.0, "epoch": 0.1085602116582743, "frac_reward_zero_std": 0.6875, "grad_norm": 0.06286207256995095, "learning_rate": 9.998160690136289e-07, "loss": 0.0199, "num_tokens": 284960489.0, "reward": 0.14453125, "reward_std": 0.13519641757011414, "rewards/accuracy_reward/mean": 0.14919355511665344, "rewards/accuracy_reward/std": 0.3566388487815857, "step": 318 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.62109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2043.0, "completions/mean_length": 1690.939453125, "completions/mean_terminated_length": 1105.6546630859375, "completions/min_length": 117.0, "completions/min_terminated_length": 117.0, "epoch": 0.1089015959716651, "frac_reward_zero_std": 0.6875, "grad_norm": 0.060823581177644775, "learning_rate": 9.998004232634777e-07, "loss": 0.0095, "num_tokens": 285897178.0, "reward": 0.123046875, "reward_std": 0.12482814490795135, "rewards/accuracy_reward/mean": 0.1270161271095276, "rewards/accuracy_reward/std": 0.33332720398902893, "step": 319 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.623046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2032.0, "completions/mean_length": 1684.44140625, "completions/mean_terminated_length": 1083.53369140625, "completions/min_length": 15.0, "completions/min_terminated_length": 15.0, "epoch": 0.1092429802850559, "frac_reward_zero_std": 0.75, "grad_norm": 0.04165122038921761, "learning_rate": 9.997841391038957e-07, "loss": 0.0166, "num_tokens": 286842028.0, "reward": 0.083984375, "reward_std": 0.0961657464504242, "rewards/accuracy_reward/mean": 0.083984375, "rewards/accuracy_reward/std": 0.2776356339454651, "step": 320 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.708984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2048.0, "completions/mean_length": 1748.01953125, "completions/mean_terminated_length": 1017.1946411132812, "completions/min_length": 207.0, "completions/min_terminated_length": 207.0, "epoch": 0.1095843645984467, "frac_reward_zero_std": 0.71875, "grad_norm": 0.04882678845907548, "learning_rate": 9.997672165579948e-07, "loss": 0.0076, "num_tokens": 287812054.0, "reward": 0.158203125, "reward_std": 0.11791393160820007, "rewards/accuracy_reward/mean": 0.158203125, "rewards/accuracy_reward/std": 0.36528825759887695, "step": 321 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.5546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2041.0, "completions/mean_length": 1667.2578125, "completions/mean_terminated_length": 1193.0, "completions/min_length": 94.0, "completions/min_terminated_length": 94.0, "epoch": 0.1099257489118375, "frac_reward_zero_std": 0.75, "grad_norm": 0.04881625353475447, "learning_rate": 9.997496556497934e-07, "loss": 0.0031, "num_tokens": 288737242.0, "reward": 0.166015625, "reward_std": 0.08864613622426987, "rewards/accuracy_reward/mean": 0.166015625, "rewards/accuracy_reward/std": 0.3724585771560669, "step": 322 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.615234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2041.0, "completions/mean_length": 1746.681640625, "completions/mean_terminated_length": 1264.878173828125, "completions/min_length": 191.0, "completions/min_terminated_length": 191.0, "epoch": 0.1102671332252283, "frac_reward_zero_std": 0.6875, "grad_norm": 0.055800776784542865, "learning_rate": 9.997314564042165e-07, "loss": 0.0042, "num_tokens": 289710151.0, "reward": 0.11328125, "reward_std": 0.13326513767242432, "rewards/accuracy_reward/mean": 0.11328125, "rewards/accuracy_reward/std": 0.3172462284564972, "step": 323 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.68359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2043.0, "completions/mean_length": 1765.369140625, "completions/mean_terminated_length": 1154.7469482421875, "completions/min_length": 173.0, "completions/min_terminated_length": 173.0, "epoch": 0.1106085175386191, "frac_reward_zero_std": 0.84375, "grad_norm": 0.036523479547995535, "learning_rate": 9.997126188470941e-07, "loss": 0.0007, "num_tokens": 290704468.0, "reward": 0.05078125, "reward_std": 0.06100328266620636, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 324 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2035.0, "completions/mean_length": 1698.853515625, "completions/mean_terminated_length": 1054.8721923828125, "completions/min_length": 220.0, "completions/min_terminated_length": 220.0, "epoch": 0.1109499018520099, "frac_reward_zero_std": 0.65625, "grad_norm": 0.06487477432683009, "learning_rate": 9.996931430051626e-07, "loss": 0.012, "num_tokens": 291643257.0, "reward": 0.208984375, "reward_std": 0.13948683440685272, "rewards/accuracy_reward/mean": 0.208984375, "rewards/accuracy_reward/std": 0.40698084235191345, "step": 325 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.68359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2036.0, "completions/mean_length": 1799.423828125, "completions/mean_terminated_length": 1262.3765869140625, "completions/min_length": 247.0, "completions/min_terminated_length": 247.0, "epoch": 0.1112912861654007, "frac_reward_zero_std": 0.6875, "grad_norm": 0.05759697721423508, "learning_rate": 9.99673028906065e-07, "loss": 0.0154, "num_tokens": 292638962.0, "reward": 0.169921875, "reward_std": 0.12576617300510406, "rewards/accuracy_reward/mean": 0.169921875, "rewards/accuracy_reward/std": 0.3759314715862274, "step": 326 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.58984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2020.0, "completions/mean_length": 1677.7421875, "completions/mean_terminated_length": 1145.2762451171875, "completions/min_length": 195.0, "completions/min_terminated_length": 195.0, "epoch": 0.1116326704787915, "frac_reward_zero_std": 0.53125, "grad_norm": 0.08341942847329462, "learning_rate": 9.996522765783488e-07, "loss": 0.0307, "num_tokens": 293572366.0, "reward": 0.28125, "reward_std": 0.20809099078178406, "rewards/accuracy_reward/mean": 0.28125, "rewards/accuracy_reward/std": 0.45004892349243164, "step": 327 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.732421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1994.0, "completions/mean_length": 1772.068359375, "completions/mean_terminated_length": 1016.781005859375, "completions/min_length": 17.0, "completions/min_terminated_length": 17.0, "epoch": 0.1119740547921823, "frac_reward_zero_std": 0.875, "grad_norm": 0.03464978667496495, "learning_rate": 9.996308860514686e-07, "loss": 0.0173, "num_tokens": 294554833.0, "reward": 0.0859375, "reward_std": 0.059578798711299896, "rewards/accuracy_reward/mean": 0.0859375, "rewards/accuracy_reward/std": 0.28054583072662354, "step": 328 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.60546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2039.0, "completions/mean_length": 1665.970703125, "completions/mean_terminated_length": 1079.6881103515625, "completions/min_length": 82.0, "completions/min_terminated_length": 82.0, "epoch": 0.1123154391055731, "frac_reward_zero_std": 0.75, "grad_norm": 0.05599199410478439, "learning_rate": 9.996088573557843e-07, "loss": 0.0047, "num_tokens": 295480034.0, "reward": 0.205078125, "reward_std": 0.08958513289690018, "rewards/accuracy_reward/mean": 0.205078125, "rewards/accuracy_reward/std": 0.4041535556316376, "step": 329 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2048.0, "completions/mean_length": 1866.75, "completions/mean_terminated_length": 1120.0, "completions/min_length": 271.0, "completions/min_terminated_length": 271.0, "epoch": 0.1126568234189639, "frac_reward_zero_std": 0.75, "grad_norm": 0.04849906900597323, "learning_rate": 9.995861905225617e-07, "loss": -0.0013, "num_tokens": 296512706.0, "reward": 0.072265625, "reward_std": 0.10177553445100784, "rewards/accuracy_reward/mean": 0.072265625, "rewards/accuracy_reward/std": 0.2591804563999176, "step": 330 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.69921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2038.0, "completions/mean_length": 1757.1953125, "completions/mean_terminated_length": 1081.1688232421875, "completions/min_length": 20.0, "completions/min_terminated_length": 20.0, "epoch": 0.1129982077323547, "frac_reward_zero_std": 0.625, "grad_norm": 0.06822229763736717, "learning_rate": 9.995628855839721e-07, "loss": 0.0124, "num_tokens": 297495030.0, "reward": 0.123046875, "reward_std": 0.14834031462669373, "rewards/accuracy_reward/mean": 0.123046875, "rewards/accuracy_reward/std": 0.32881227135658264, "step": 331 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2028.0, "completions/mean_length": 1781.4609375, "completions/mean_terminated_length": 1100.3055419921875, "completions/min_length": 120.0, "completions/min_terminated_length": 120.0, "epoch": 0.1133395920457455, "frac_reward_zero_std": 0.59375, "grad_norm": 0.09081261286430949, "learning_rate": 9.995389425730923e-07, "loss": 0.0269, "num_tokens": 298486482.0, "reward": 0.142578125, "reward_std": 0.1532699465751648, "rewards/accuracy_reward/mean": 0.142578125, "rewards/accuracy_reward/std": 0.3499840497970581, "step": 332 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.62109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2036.0, "completions/mean_length": 1763.40234375, "completions/mean_terminated_length": 1296.8968505859375, "completions/min_length": 14.0, "completions/min_terminated_length": 14.0, "epoch": 0.1136809763591363, "frac_reward_zero_std": 0.625, "grad_norm": 0.07955616710976351, "learning_rate": 9.995143615239056e-07, "loss": 0.0092, "num_tokens": 299467408.0, "reward": 0.130859375, "reward_std": 0.1380920112133026, "rewards/accuracy_reward/mean": 0.130859375, "rewards/accuracy_reward/std": 0.33757632970809937, "step": 333 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.603515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2048.0, "completions/mean_length": 1673.837890625, "completions/mean_terminated_length": 1104.3004150390625, "completions/min_length": 28.0, "completions/min_terminated_length": 28.0, "epoch": 0.11402236067252709, "frac_reward_zero_std": 0.625, "grad_norm": 0.08875806539926845, "learning_rate": 9.994891424712998e-07, "loss": 0.0109, "num_tokens": 300401645.0, "reward": 0.14453125, "reward_std": 0.1489243358373642, "rewards/accuracy_reward/mean": 0.14453125, "rewards/accuracy_reward/std": 0.35197147727012634, "step": 334 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.603515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2029.0, "completions/mean_length": 1626.884765625, "completions/mean_terminated_length": 985.8768310546875, "completions/min_length": 83.0, "completions/min_terminated_length": 83.0, "epoch": 0.11436374498591789, "frac_reward_zero_std": 0.59375, "grad_norm": 0.10709636875964579, "learning_rate": 9.99463285451069e-07, "loss": 0.0137, "num_tokens": 301311618.0, "reward": 0.16796875, "reward_std": 0.16807593405246735, "rewards/accuracy_reward/mean": 0.16796875, "rewards/accuracy_reward/std": 0.374204158782959, "step": 335 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2033.0, "completions/mean_length": 1716.482421875, "completions/mean_terminated_length": 1105.0167236328125, "completions/min_length": 185.0, "completions/min_terminated_length": 185.0, "epoch": 0.11470512929930869, "frac_reward_zero_std": 0.5625, "grad_norm": 0.0704848028686384, "learning_rate": 9.994367904999127e-07, "loss": 0.008, "num_tokens": 302264345.0, "reward": 0.216796875, "reward_std": 0.1845674067735672, "rewards/accuracy_reward/mean": 0.216796875, "rewards/accuracy_reward/std": 0.4124660789966583, "step": 336 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.67578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2044.0, "completions/mean_length": 1739.603515625, "completions/mean_terminated_length": 1096.8011474609375, "completions/min_length": 26.0, "completions/min_terminated_length": 26.0, "epoch": 0.11504651361269949, "frac_reward_zero_std": 0.65625, "grad_norm": 0.09842404712020399, "learning_rate": 9.994096576554353e-07, "loss": 0.009, "num_tokens": 303231742.0, "reward": 0.162109375, "reward_std": 0.13342025876045227, "rewards/accuracy_reward/mean": 0.16733871400356293, "rewards/accuracy_reward/std": 0.37365487217903137, "step": 337 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.658203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2045.0, "completions/mean_length": 1731.248046875, "completions/mean_terminated_length": 1121.5313720703125, "completions/min_length": 40.0, "completions/min_terminated_length": 40.0, "epoch": 0.11538789792609029, "frac_reward_zero_std": 0.65625, "grad_norm": 0.08891326086421894, "learning_rate": 9.993818869561467e-07, "loss": 0.0155, "num_tokens": 304202029.0, "reward": 0.080078125, "reward_std": 0.11685626953840256, "rewards/accuracy_reward/mean": 0.080078125, "rewards/accuracy_reward/std": 0.271679550409317, "step": 338 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.658203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2034.0, "completions/mean_length": 1747.025390625, "completions/mean_terminated_length": 1167.434326171875, "completions/min_length": 183.0, "completions/min_terminated_length": 183.0, "epoch": 0.11572928223948109, "frac_reward_zero_std": 0.59375, "grad_norm": 0.09529601905784285, "learning_rate": 9.99353478441463e-07, "loss": 0.0143, "num_tokens": 305178042.0, "reward": 0.162109375, "reward_std": 0.17205162346363068, "rewards/accuracy_reward/mean": 0.162109375, "rewards/accuracy_reward/std": 0.3689115643501282, "step": 339 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.66796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.0, "completions/mean_length": 1745.998046875, "completions/mean_terminated_length": 1138.441162109375, "completions/min_length": 50.0, "completions/min_terminated_length": 50.0, "epoch": 0.11607066655287189, "frac_reward_zero_std": 0.71875, "grad_norm": 0.07301399070834347, "learning_rate": 9.993244321517045e-07, "loss": 0.0105, "num_tokens": 306143929.0, "reward": 0.107421875, "reward_std": 0.11646249890327454, "rewards/accuracy_reward/mean": 0.11088709533214569, "rewards/accuracy_reward/std": 0.3143092691898346, "step": 340 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.662109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2042.0, "completions/mean_length": 1731.720703125, "completions/mean_terminated_length": 1112.24853515625, "completions/min_length": 149.0, "completions/min_terminated_length": 149.0, "epoch": 0.11641205086626269, "frac_reward_zero_std": 0.71875, "grad_norm": 0.08293484324941816, "learning_rate": 9.99294748128097e-07, "loss": 0.0134, "num_tokens": 307108522.0, "reward": 0.162109375, "reward_std": 0.10481776297092438, "rewards/accuracy_reward/mean": 0.162109375, "rewards/accuracy_reward/std": 0.3689115643501282, "step": 341 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2048.0, "completions/mean_length": 1765.857421875, "completions/mean_terminated_length": 1145.4625244140625, "completions/min_length": 24.0, "completions/min_terminated_length": 24.0, "epoch": 0.11675343517965349, "frac_reward_zero_std": 0.71875, "grad_norm": 0.13179741631114522, "learning_rate": 9.992644264127717e-07, "loss": 0.0216, "num_tokens": 308086321.0, "reward": 0.12109375, "reward_std": 0.10998040437698364, "rewards/accuracy_reward/mean": 0.125, "rewards/accuracy_reward/std": 0.3310528099536896, "step": 342 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.724609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2048.0, "completions/mean_length": 1842.978515625, "completions/mean_terminated_length": 1303.5247802734375, "completions/min_length": 64.0, "completions/min_terminated_length": 64.0, "epoch": 0.11709481949304429, "frac_reward_zero_std": 0.78125, "grad_norm": 0.07285559281916934, "learning_rate": 9.992334670487646e-07, "loss": 0.0089, "num_tokens": 309098662.0, "reward": 0.068359375, "reward_std": 0.08750899136066437, "rewards/accuracy_reward/mean": 0.068359375, "rewards/accuracy_reward/std": 0.25260838866233826, "step": 343 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.751953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.0, "completions/mean_length": 1880.46484375, "completions/mean_terminated_length": 1372.5826416015625, "completions/min_length": 16.0, "completions/min_terminated_length": 16.0, "epoch": 0.11743620380643509, "frac_reward_zero_std": 0.78125, "grad_norm": 0.08848032353536818, "learning_rate": 9.99201870080017e-07, "loss": 0.006, "num_tokens": 310137524.0, "reward": 0.12109375, "reward_std": 0.09862300753593445, "rewards/accuracy_reward/mean": 0.12109375, "rewards/accuracy_reward/std": 0.3265552520751953, "step": 344 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.75390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2038.0, "completions/mean_length": 1813.697265625, "completions/mean_terminated_length": 1095.9127197265625, "completions/min_length": 41.0, "completions/min_terminated_length": 41.0, "epoch": 0.11777758811982589, "frac_reward_zero_std": 0.75, "grad_norm": 0.09133029827215806, "learning_rate": 9.99169635551375e-07, "loss": -0.0024, "num_tokens": 311145289.0, "reward": 0.07421875, "reward_std": 0.10195636004209518, "rewards/accuracy_reward/mean": 0.07421875, "rewards/accuracy_reward/std": 0.2623828947544098, "step": 345 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.775390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1896.638671875, "completions/mean_terminated_length": 1374.113037109375, "completions/min_length": 62.0, "completions/min_terminated_length": 62.0, "epoch": 0.11811897243321669, "frac_reward_zero_std": 0.78125, "grad_norm": 0.13596286673083904, "learning_rate": 9.991367635085897e-07, "loss": 0.0078, "num_tokens": 312199584.0, "reward": 0.103515625, "reward_std": 0.07834453880786896, "rewards/accuracy_reward/mean": 0.103515625, "rewards/accuracy_reward/std": 0.30492907762527466, "step": 346 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.67578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1924.0, "completions/mean_length": 1679.658203125, "completions/mean_terminated_length": 911.9096069335938, "completions/min_length": 18.0, "completions/min_terminated_length": 18.0, "epoch": 0.11846035674660749, "frac_reward_zero_std": 0.75, "grad_norm": 0.1534179772580519, "learning_rate": 9.991032539983166e-07, "loss": 0.0148, "num_tokens": 313133617.0, "reward": 0.150390625, "reward_std": 0.1024552434682846, "rewards/accuracy_reward/mean": 0.150390625, "rewards/accuracy_reward/std": 0.35780346393585205, "step": 347 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1935.0, "completions/mean_length": 1832.822265625, "completions/mean_terminated_length": 1159.524169921875, "completions/min_length": 103.0, "completions/min_terminated_length": 103.0, "epoch": 0.11880174105999829, "frac_reward_zero_std": 0.625, "grad_norm": 0.1596366441630623, "learning_rate": 9.990691070681169e-07, "loss": 0.0208, "num_tokens": 314149462.0, "reward": 0.1875, "reward_std": 0.15019595623016357, "rewards/accuracy_reward/mean": 0.1875, "rewards/accuracy_reward/std": 0.39069411158561707, "step": 348 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.69921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2039.0, "completions/mean_length": 1768.361328125, "completions/mean_terminated_length": 1118.292236328125, "completions/min_length": 27.0, "completions/min_terminated_length": 27.0, "epoch": 0.11914312537338909, "frac_reward_zero_std": 0.75, "grad_norm": 0.11695111442667908, "learning_rate": 9.990343227664552e-07, "loss": 0.0099, "num_tokens": 315132079.0, "reward": 0.109375, "reward_std": 0.10326766967773438, "rewards/accuracy_reward/mean": 0.109375, "rewards/accuracy_reward/std": 0.31241437792778015, "step": 349 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.818359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2028.0, "completions/mean_length": 1922.390625, "completions/mean_terminated_length": 1356.47314453125, "completions/min_length": 21.0, "completions/min_terminated_length": 21.0, "epoch": 0.11948450968677989, "frac_reward_zero_std": 0.84375, "grad_norm": 0.12473908865500574, "learning_rate": 9.98998901142702e-07, "loss": 0.0044, "num_tokens": 316187159.0, "reward": 0.033203125, "reward_std": 0.06166848540306091, "rewards/accuracy_reward/mean": 0.033203125, "rewards/accuracy_reward/std": 0.17934183776378632, "step": 350 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.748046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2042.0, "completions/mean_length": 1816.8984375, "completions/mean_terminated_length": 1130.7596435546875, "completions/min_length": 153.0, "completions/min_terminated_length": 153.0, "epoch": 0.1198258940001707, "frac_reward_zero_std": 0.75, "grad_norm": 0.11734946077064011, "learning_rate": 9.989628422471316e-07, "loss": 0.0125, "num_tokens": 317196627.0, "reward": 0.12890625, "reward_std": 0.10067909955978394, "rewards/accuracy_reward/mean": 0.13306452333927155, "rewards/accuracy_reward/std": 0.3399873375892639, "step": 351 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.80078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2042.0, "completions/mean_length": 1847.251953125, "completions/mean_terminated_length": 1040.3236083984375, "completions/min_length": 17.0, "completions/min_terminated_length": 17.0, "epoch": 0.1201672783135615, "frac_reward_zero_std": 0.84375, "grad_norm": 0.14179644403465022, "learning_rate": 9.989261461309232e-07, "loss": 0.0019, "num_tokens": 318209540.0, "reward": 0.041015625, "reward_std": 0.048086829483509064, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 352 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.763671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2042.0, "completions/mean_length": 1842.19921875, "completions/mean_terminated_length": 1177.1734619140625, "completions/min_length": 16.0, "completions/min_terminated_length": 16.0, "epoch": 0.1205086626269523, "frac_reward_zero_std": 0.78125, "grad_norm": 0.2728144471324912, "learning_rate": 9.9888881284616e-07, "loss": 0.0144, "num_tokens": 319235594.0, "reward": 0.091796875, "reward_std": 0.07421942055225372, "rewards/accuracy_reward/mean": 0.091796875, "rewards/accuracy_reward/std": 0.289021372795105, "step": 353 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.841796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2027.0, "completions/mean_length": 1896.412109375, "completions/mean_terminated_length": 1089.8148193359375, "completions/min_length": 199.0, "completions/min_terminated_length": 199.0, "epoch": 0.1208500469403431, "frac_reward_zero_std": 0.90625, "grad_norm": 0.0710082246771826, "learning_rate": 9.9885084244583e-07, "loss": 0.0032, "num_tokens": 320288061.0, "reward": 0.044921875, "reward_std": 0.030584799125790596, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 354 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2021.0, "completions/mean_length": 1872.99609375, "completions/mean_terminated_length": 1275.5689697265625, "completions/min_length": 20.0, "completions/min_terminated_length": 20.0, "epoch": 0.1211914312537339, "frac_reward_zero_std": 0.8125, "grad_norm": 0.9755373756315711, "learning_rate": 9.988122349838247e-07, "loss": 0.007, "num_tokens": 321321899.0, "reward": 0.09765625, "reward_std": 0.06233368441462517, "rewards/accuracy_reward/mean": 0.10080645233392715, "rewards/accuracy_reward/std": 0.30137622356414795, "step": 355 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2044.0, "completions/mean_length": 1834.01171875, "completions/mean_terminated_length": 994.519287109375, "completions/min_length": 27.0, "completions/min_terminated_length": 27.0, "epoch": 0.1215328155671247, "frac_reward_zero_std": 0.875, "grad_norm": 1.0202236434349419, "learning_rate": 9.987729905149411e-07, "loss": 0.0071, "num_tokens": 322341505.0, "reward": 0.06640625, "reward_std": 0.049843885004520416, "rewards/accuracy_reward/mean": 0.06640625, "rewards/accuracy_reward/std": 0.2492343932390213, "step": 356 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.732421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2023.0, "completions/mean_length": 1780.455078125, "completions/mean_terminated_length": 1048.1240234375, "completions/min_length": 30.0, "completions/min_terminated_length": 30.0, "epoch": 0.1218741998805155, "frac_reward_zero_std": 0.75, "grad_norm": 1.3219660564537818, "learning_rate": 9.98733109094879e-07, "loss": 0.0052, "num_tokens": 323329162.0, "reward": 0.09765625, "reward_std": 0.09490203112363815, "rewards/accuracy_reward/mean": 0.09765625, "rewards/accuracy_reward/std": 0.29713961482048035, "step": 357 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.724609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2042.0, "completions/mean_length": 1752.212890625, "completions/mean_terminated_length": 973.9361572265625, "completions/min_length": 78.0, "completions/min_terminated_length": 78.0, "epoch": 0.1222155841939063, "frac_reward_zero_std": 0.6875, "grad_norm": 1.8425893010360381, "learning_rate": 9.98692590780243e-07, "loss": 0.0093, "num_tokens": 324317063.0, "reward": 0.09765625, "reward_std": 0.13064473867416382, "rewards/accuracy_reward/mean": 0.09765625, "rewards/accuracy_reward/std": 0.29713961482048035, "step": 358 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.845703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2004.0, "completions/mean_length": 1910.25, "completions/mean_terminated_length": 1155.240478515625, "completions/min_length": 37.0, "completions/min_terminated_length": 37.0, "epoch": 0.1225569685072971, "frac_reward_zero_std": 0.8125, "grad_norm": 2.3807639689777997, "learning_rate": 9.986514356285412e-07, "loss": 0.0067, "num_tokens": 325376983.0, "reward": 0.0390625, "reward_std": 0.07241567224264145, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 359 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.79296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2041.0, "completions/mean_length": 1847.900390625, "completions/mean_terminated_length": 1081.481201171875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1228983528206879, "frac_reward_zero_std": 0.84375, "grad_norm": 3.4733312131034353, "learning_rate": 9.986096436981862e-07, "loss": 0.0118, "num_tokens": 326405972.0, "reward": 0.04296875, "reward_std": 0.0635918527841568, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 360 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.740234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2020.0, "completions/mean_length": 1812.865234375, "completions/mean_terminated_length": 1142.819580078125, "completions/min_length": 112.0, "completions/min_terminated_length": 112.0, "epoch": 0.1232397371340787, "frac_reward_zero_std": 0.78125, "grad_norm": 3.6548073897780564, "learning_rate": 9.985672150484937e-07, "loss": 0.0128, "num_tokens": 327415983.0, "reward": 0.16015625, "reward_std": 0.08980467915534973, "rewards/accuracy_reward/mean": 0.16015625, "rewards/accuracy_reward/std": 0.3671095669269562, "step": 361 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.69921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.0, "completions/mean_length": 1768.724609375, "completions/mean_terminated_length": 1119.5, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1235811214474695, "frac_reward_zero_std": 0.78125, "grad_norm": 3.888454704512162, "learning_rate": 9.985241497396835e-07, "loss": -0.004, "num_tokens": 328403922.0, "reward": 0.140625, "reward_std": 0.08251741528511047, "rewards/accuracy_reward/mean": 0.140625, "rewards/accuracy_reward/std": 0.3479743003845215, "step": 362 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.630859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2042.0, "completions/mean_length": 1630.876953125, "completions/mean_terminated_length": 918.3333129882812, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1239225057608603, "frac_reward_zero_std": 0.65625, "grad_norm": 6.702267300913001, "learning_rate": 9.984804478328792e-07, "loss": 0.0061, "num_tokens": 329321331.0, "reward": 0.1328125, "reward_std": 0.14254970848560333, "rewards/accuracy_reward/mean": 0.1328125, "rewards/accuracy_reward/std": 0.33970388770103455, "step": 363 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.861328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2016.0, "completions/mean_length": 1916.169921875, "completions/mean_terminated_length": 1097.3380126953125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1242638900742511, "frac_reward_zero_std": 0.875, "grad_norm": 2.1279630623853594, "learning_rate": 9.98436109390107e-07, "loss": 0.0024, "num_tokens": 330369898.0, "reward": 0.04296875, "reward_std": 0.04175759106874466, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 364 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2010.0, "completions/mean_length": 1886.998046875, "completions/mean_terminated_length": 1189.322998046875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.12460527438764189, "frac_reward_zero_std": 0.6875, "grad_norm": 4.009799030492814, "learning_rate": 9.983911344742979e-07, "loss": 0.0181, "num_tokens": 331411833.0, "reward": 0.08203125, "reward_std": 0.12440948933362961, "rewards/accuracy_reward/mean": 0.08203125, "rewards/accuracy_reward/std": 0.2746807038784027, "step": 365 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.806640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2040.0, "completions/mean_length": 1858.69140625, "completions/mean_terminated_length": 1068.949462890625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.12494665870103269, "frac_reward_zero_std": 0.6875, "grad_norm": 4.908216668332833, "learning_rate": 9.983455231492852e-07, "loss": 0.0051, "num_tokens": 332440283.0, "reward": 0.078125, "reward_std": 0.11812658607959747, "rewards/accuracy_reward/mean": 0.078125, "rewards/accuracy_reward/std": 0.26863065361976624, "step": 366 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2033.0, "completions/mean_length": 1769.083984375, "completions/mean_terminated_length": 896.3467407226562, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.12528804301442348, "frac_reward_zero_std": 0.8125, "grad_norm": 4.478707134866254, "learning_rate": 9.982992754798057e-07, "loss": 0.005, "num_tokens": 333423414.0, "reward": 0.076171875, "reward_std": 0.06849660724401474, "rewards/accuracy_reward/mean": 0.076171875, "rewards/accuracy_reward/std": 0.26553234457969666, "step": 367 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.732421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1983.0, "completions/mean_length": 1798.90234375, "completions/mean_terminated_length": 1117.065673828125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1256294273278143, "frac_reward_zero_std": 0.625, "grad_norm": 11.054539392413355, "learning_rate": 9.982523915315e-07, "loss": 0.0116, "num_tokens": 334416772.0, "reward": 0.1328125, "reward_std": 0.14584887027740479, "rewards/accuracy_reward/mean": 0.1328125, "rewards/accuracy_reward/std": 0.33970388770103455, "step": 368 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.802734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1965.0, "completions/mean_length": 1816.171875, "completions/mean_terminated_length": 872.7920532226562, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.12597081164120508, "frac_reward_zero_std": 0.75, "grad_norm": 9.805763736629649, "learning_rate": 9.982048713709109e-07, "loss": 0.0089, "num_tokens": 335419196.0, "reward": 0.044921875, "reward_std": 0.09032991528511047, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 369 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.787109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2028.0, "completions/mean_length": 1882.212890625, "completions/mean_terminated_length": 1269.2568359375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1263121959545959, "frac_reward_zero_std": 0.875, "grad_norm": 4.655856691146327, "learning_rate": 9.981567150654848e-07, "loss": 0.0042, "num_tokens": 336458841.0, "reward": 0.064453125, "reward_std": 0.05529290437698364, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 370 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.78515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.0, "completions/mean_length": 1858.826171875, "completions/mean_terminated_length": 1167.4818115234375, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.12665358026798668, "frac_reward_zero_std": 0.8125, "grad_norm": 12.619375394965113, "learning_rate": 9.981079226835704e-07, "loss": 0.0109, "num_tokens": 337487216.0, "reward": 0.037109375, "reward_std": 0.07026061415672302, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 371 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.90234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2035.0, "completions/mean_length": 1955.298828125, "completions/mean_terminated_length": 1098.739990234375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1269949645813775, "frac_reward_zero_std": 0.9375, "grad_norm": 3.543135339261255, "learning_rate": 9.9805849429442e-07, "loss": -0.0013, "num_tokens": 338579353.0, "reward": 0.0234375, "reward_std": 0.028608137741684914, "rewards/accuracy_reward/mean": 0.0234375, "rewards/accuracy_reward/std": 0.15143637359142303, "step": 372 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.888671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1956.0, "completions/mean_length": 1904.044921875, "completions/mean_terminated_length": 754.9298095703125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.12733634889476828, "frac_reward_zero_std": 0.84375, "grad_norm": 11.115825424624719, "learning_rate": 9.980084299681879e-07, "loss": 0.01, "num_tokens": 339628848.0, "reward": 0.04296875, "reward_std": 0.050948236137628555, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 373 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1992.0, "completions/mean_length": 1923.17578125, "completions/mean_terminated_length": 1108.1470947265625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1276777332081591, "frac_reward_zero_std": 0.84375, "grad_norm": 9.997232530856346, "learning_rate": 9.979577297759312e-07, "loss": 0.001, "num_tokens": 340687722.0, "reward": 0.048828125, "reward_std": 0.05138834938406944, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 374 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.794921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2006.0, "completions/mean_length": 1859.162109375, "completions/mean_terminated_length": 1127.1905517578125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.12801911752154987, "frac_reward_zero_std": 0.75, "grad_norm": 15.97417048651692, "learning_rate": 9.979063937896098e-07, "loss": 0.0279, "num_tokens": 341721853.0, "reward": 0.099609375, "reward_std": 0.09506139904260635, "rewards/accuracy_reward/mean": 0.10282257944345474, "rewards/accuracy_reward/std": 0.30403366684913635, "step": 375 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.802734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2042.0, "completions/mean_length": 1903.326171875, "completions/mean_terminated_length": 1315.019775390625, "completions/min_length": 7.0, "completions/min_terminated_length": 7.0, "epoch": 0.1283605018349407, "frac_reward_zero_std": 0.8125, "grad_norm": 6.2069615640212055, "learning_rate": 9.978544220820858e-07, "loss": 0.004, "num_tokens": 342776836.0, "reward": 0.095703125, "reward_std": 0.0799964889883995, "rewards/accuracy_reward/mean": 0.095703125, "rewards/accuracy_reward/std": 0.2944713830947876, "step": 376 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1986.0, "completions/mean_length": 1892.61328125, "completions/mean_terminated_length": 943.0277709960938, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.12870188614833147, "frac_reward_zero_std": 0.90625, "grad_norm": 6.869596712335641, "learning_rate": 9.978018147271232e-07, "loss": 0.0085, "num_tokens": 343813502.0, "reward": 0.009765625, "reward_std": 0.029160313308238983, "rewards/accuracy_reward/mean": 0.010080644860863686, "rewards/accuracy_reward/std": 0.0999959409236908, "step": 377 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.9140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2030.0, "completions/mean_length": 1964.162109375, "completions/mean_terminated_length": 1072.431884765625, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.1290432704617223, "frac_reward_zero_std": 0.9375, "grad_norm": 4.789800421844635, "learning_rate": 9.977485717993885e-07, "loss": 0.0018, "num_tokens": 344888625.0, "reward": 0.0078125, "reward_std": 0.020409777760505676, "rewards/accuracy_reward/mean": 0.0078125, "rewards/accuracy_reward/std": 0.08812850713729858, "step": 378 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.93359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1954.0, "completions/mean_length": 1973.662109375, "completions/mean_terminated_length": 928.558837890625, "completions/min_length": 11.0, "completions/min_terminated_length": 11.0, "epoch": 0.12938465477511307, "frac_reward_zero_std": 0.9375, "grad_norm": 4.103245270166282, "learning_rate": 9.976946933744505e-07, "loss": 0.0027, "num_tokens": 345974036.0, "reward": 0.005859375, "reward_std": 0.01848640665411949, "rewards/accuracy_reward/mean": 0.005859375, "rewards/accuracy_reward/std": 0.07639661431312561, "step": 379 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.87109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2039.0, "completions/mean_length": 1924.02734375, "completions/mean_terminated_length": 1086.272705078125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.12972603908850389, "frac_reward_zero_std": 0.875, "grad_norm": 5.777409629720089, "learning_rate": 9.9764017952878e-07, "loss": 0.005, "num_tokens": 347034722.0, "reward": 0.033203125, "reward_std": 0.04219770431518555, "rewards/accuracy_reward/mean": 0.033203125, "rewards/accuracy_reward/std": 0.17934183776378632, "step": 380 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.87890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1977.0, "completions/mean_length": 1920.111328125, "completions/mean_terminated_length": 991.8870849609375, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.13006742340189467, "frac_reward_zero_std": 0.9375, "grad_norm": 8.452264719758775, "learning_rate": 9.97585030339749e-07, "loss": 0.0053, "num_tokens": 348095563.0, "reward": 0.029296875, "reward_std": 0.026572702452540398, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 381 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2000.0, "completions/mean_length": 1943.615234375, "completions/mean_terminated_length": 1157.25, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.13040880771528549, "frac_reward_zero_std": 0.8125, "grad_norm": 12.257495929369215, "learning_rate": 9.97529245885632e-07, "loss": 0.0048, "num_tokens": 349164470.0, "reward": 0.041015625, "reward_std": 0.059305962175130844, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 382 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.826171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2044.0, "completions/mean_length": 1895.861328125, "completions/mean_terminated_length": 1173.2696533203125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.13075019202867627, "frac_reward_zero_std": 0.84375, "grad_norm": 12.64986476935335, "learning_rate": 9.974728262456043e-07, "loss": -0.0002, "num_tokens": 350214127.0, "reward": 0.0546875, "reward_std": 0.06062985956668854, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 383 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.88671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2026.0, "completions/mean_length": 1943.9765625, "completions/mean_terminated_length": 1129.72412109375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.13109157634206708, "frac_reward_zero_std": 0.875, "grad_norm": 14.543445015035747, "learning_rate": 9.974157714997438e-07, "loss": 0.003, "num_tokens": 351281411.0, "reward": 0.01953125, "reward_std": 0.04505910724401474, "rewards/accuracy_reward/mean": 0.01953125, "rewards/accuracy_reward/std": 0.1385180652141571, "step": 384 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.85546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2044.0, "completions/mean_length": 1888.806640625, "completions/mean_terminated_length": 947.4054565429688, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.13143296065545787, "frac_reward_zero_std": 0.875, "grad_norm": 7.967585905732063, "learning_rate": 9.973580817290295e-07, "loss": 0.0074, "num_tokens": 352333008.0, "reward": 0.015625, "reward_std": 0.04081955552101135, "rewards/accuracy_reward/mean": 0.016129031777381897, "rewards/accuracy_reward/std": 0.12609896063804626, "step": 385 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2023.0, "completions/mean_length": 1891.8203125, "completions/mean_terminated_length": 937.388916015625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.13177434496884868, "frac_reward_zero_std": 0.78125, "grad_norm": 17.782230088734398, "learning_rate": 9.972997570153406e-07, "loss": 0.0132, "num_tokens": 353372308.0, "reward": 0.056640625, "reward_std": 0.07521171122789383, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 386 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1967.68359375, "completions/mean_terminated_length": 1313.6785888671875, "completions/min_length": 421.0, "completions/min_terminated_length": 421.0, "epoch": 0.13211572928223947, "frac_reward_zero_std": 1.0, "grad_norm": 0.0, "learning_rate": 9.972407974414592e-07, "loss": 0.0, "num_tokens": 354461682.0, "reward": 0.0, "reward_std": 0.0, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "step": 387 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.849609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2032.0, "completions/mean_length": 1897.298828125, "completions/mean_terminated_length": 1046.727294921875, "completions/min_length": 15.0, "completions/min_terminated_length": 15.0, "epoch": 0.13245711359563028, "frac_reward_zero_std": 0.875, "grad_norm": 14.200278918139587, "learning_rate": 9.971812030910671e-07, "loss": 0.0, "num_tokens": 355521803.0, "reward": 0.037109375, "reward_std": 0.049345001578330994, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 388 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2043.0, "completions/mean_length": 1897.673828125, "completions/mean_terminated_length": 979.013916015625, "completions/min_length": 75.0, "completions/min_terminated_length": 75.0, "epoch": 0.13279849790902107, "frac_reward_zero_std": 0.96875, "grad_norm": 9.098451304623806, "learning_rate": 9.971209740487478e-07, "loss": -0.0004, "num_tokens": 356573508.0, "reward": 0.001953125, "reward_std": 0.0078125, "rewards/accuracy_reward/mean": 0.001953125, "rewards/accuracy_reward/std": 0.04419417306780815, "step": 389 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.84765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1998.0, "completions/mean_length": 1918.1484375, "completions/mean_terminated_length": 1195.6409912109375, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.13313988222241188, "frac_reward_zero_std": 0.90625, "grad_norm": 8.719509731807541, "learning_rate": 9.970601103999854e-07, "loss": 0.002, "num_tokens": 357632784.0, "reward": 0.052734375, "reward_std": 0.04456022381782532, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 390 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.900390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1979.0, "completions/mean_length": 1968.681640625, "completions/mean_terminated_length": 1251.7059326171875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.13348126653580267, "frac_reward_zero_std": 0.8125, "grad_norm": 8.949437446080502, "learning_rate": 9.969986122311648e-07, "loss": 0.0117, "num_tokens": 358719645.0, "reward": 0.02734375, "reward_std": 0.06233368441462517, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 391 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.94140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1981.0, "completions/mean_length": 1994.478515625, "completions/mean_terminated_length": 1134.5667724609375, "completions/min_length": 20.0, "completions/min_terminated_length": 20.0, "epoch": 0.13382265084919348, "frac_reward_zero_std": 0.90625, "grad_norm": 7.397498252199284, "learning_rate": 9.969364796295712e-07, "loss": 0.008, "num_tokens": 359814194.0, "reward": 0.01171875, "reward_std": 0.02960042469203472, "rewards/accuracy_reward/mean": 0.01171875, "rewards/accuracy_reward/std": 0.10772226005792618, "step": 392 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.9140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1880.0, "completions/mean_length": 1957.96484375, "completions/mean_terminated_length": 1000.3182373046875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.13416403516258427, "frac_reward_zero_std": 0.8125, "grad_norm": 14.766023961661254, "learning_rate": 9.968737126833905e-07, "loss": 0.0124, "num_tokens": 360892096.0, "reward": 0.056640625, "reward_std": 0.07438573241233826, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 393 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.91015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1990.0, "completions/mean_length": 1976.1875, "completions/mean_terminated_length": 1248.6956787109375, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.13450541947597508, "frac_reward_zero_std": 0.9375, "grad_norm": 2.7653090694364484, "learning_rate": 9.96810311481709e-07, "loss": 0.0002, "num_tokens": 361973632.0, "reward": 0.01171875, "reward_std": 0.022772299125790596, "rewards/accuracy_reward/mean": 0.01171875, "rewards/accuracy_reward/std": 0.10772226005792618, "step": 394 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.830078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1978.0, "completions/mean_length": 1907.5234375, "completions/mean_terminated_length": 1221.287353515625, "completions/min_length": 7.0, "completions/min_terminated_length": 7.0, "epoch": 0.13484680378936587, "frac_reward_zero_std": 0.8125, "grad_norm": 30.105814068991844, "learning_rate": 9.967462761145132e-07, "loss": 0.0206, "num_tokens": 363019852.0, "reward": 0.08203125, "reward_std": 0.0697084441781044, "rewards/accuracy_reward/mean": 0.08203125, "rewards/accuracy_reward/std": 0.2746807038784027, "step": 395 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.892578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1894.0, "completions/mean_length": 1937.1484375, "completions/mean_terminated_length": 1016.0726928710938, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.13518818810275668, "frac_reward_zero_std": 0.875, "grad_norm": 12.441591806308416, "learning_rate": 9.966816066726895e-07, "loss": 0.0057, "num_tokens": 364092056.0, "reward": 0.0234375, "reward_std": 0.049843885004520416, "rewards/accuracy_reward/mean": 0.0234375, "rewards/accuracy_reward/std": 0.15143637359142303, "step": 396 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.873046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2024.0, "completions/mean_length": 1932.779296875, "completions/mean_terminated_length": 1140.4154052734375, "completions/min_length": 461.0, "completions/min_terminated_length": 461.0, "epoch": 0.13552957241614746, "frac_reward_zero_std": 0.84375, "grad_norm": 8.202296725686962, "learning_rate": 9.96616303248024e-07, "loss": 0.0059, "num_tokens": 365152743.0, "reward": 0.046875, "reward_std": 0.06751786917448044, "rewards/accuracy_reward/mean": 0.046875, "rewards/accuracy_reward/std": 0.21157780289649963, "step": 397 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2030.0, "completions/mean_length": 1993.341796875, "completions/mean_terminated_length": 1348.375, "completions/min_length": 474.0, "completions/min_terminated_length": 474.0, "epoch": 0.13587095672953828, "frac_reward_zero_std": 0.9375, "grad_norm": 5.469273864133791, "learning_rate": 9.965503659332037e-07, "loss": 0.0093, "num_tokens": 366247110.0, "reward": 0.005859375, "reward_std": 0.01848640665411949, "rewards/accuracy_reward/mean": 0.005859375, "rewards/accuracy_reward/std": 0.07639661431312561, "step": 398 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.892578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.0, "completions/mean_length": 1951.3359375, "completions/mean_terminated_length": 1148.1453857421875, "completions/min_length": 86.0, "completions/min_terminated_length": 86.0, "epoch": 0.13621234104292906, "frac_reward_zero_std": 0.71875, "grad_norm": 15.228366242907972, "learning_rate": 9.964837948218138e-07, "loss": 0.0213, "num_tokens": 367328210.0, "reward": 0.08984375, "reward_std": 0.1060623973608017, "rewards/accuracy_reward/mean": 0.08984375, "rewards/accuracy_reward/std": 0.2862374484539032, "step": 399 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.888671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2000.0, "completions/mean_length": 1942.00390625, "completions/mean_terminated_length": 1095.894775390625, "completions/min_length": 357.0, "completions/min_terminated_length": 357.0, "epoch": 0.13655372535631988, "frac_reward_zero_std": 0.90625, "grad_norm": 10.869894102290171, "learning_rate": 9.964165900083402e-07, "loss": 0.0082, "num_tokens": 368406772.0, "reward": 0.0234375, "reward_std": 0.036547206342220306, "rewards/accuracy_reward/mean": 0.0234375, "rewards/accuracy_reward/std": 0.15143637359142303, "step": 400 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.82421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2041.0, "completions/mean_length": 1893.1640625, "completions/mean_terminated_length": 1167.7667236328125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1368951096697107, "frac_reward_zero_std": 0.84375, "grad_norm": 5.98695375449312, "learning_rate": 9.963487515881678e-07, "loss": 0.0023, "num_tokens": 369460120.0, "reward": 0.056640625, "reward_std": 0.05628518760204315, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 401 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.962890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1976.0, "completions/mean_length": 2008.505859375, "completions/mean_terminated_length": 983.7368774414062, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.13723649398310148, "frac_reward_zero_std": 0.9375, "grad_norm": 10.270718080635893, "learning_rate": 9.962802796575811e-07, "loss": 0.0082, "num_tokens": 370553259.0, "reward": 0.01171875, "reward_std": 0.025194555521011353, "rewards/accuracy_reward/mean": 0.01171875, "rewards/accuracy_reward/std": 0.10772226005792618, "step": 402 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1894.0, "completions/mean_length": 1902.876953125, "completions/mean_terminated_length": 956.0294189453125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1375778782964923, "frac_reward_zero_std": 0.75, "grad_norm": 10.95570444122586, "learning_rate": 9.96211174313763e-07, "loss": -0.0004, "num_tokens": 371611740.0, "reward": 0.060546875, "reward_std": 0.0870293527841568, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 403 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.857421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1998.0, "completions/mean_length": 1904.462890625, "completions/mean_terminated_length": 1041.27392578125, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "epoch": 0.13791926260988308, "frac_reward_zero_std": 0.84375, "grad_norm": 11.238549811352238, "learning_rate": 9.961414356547962e-07, "loss": 0.0059, "num_tokens": 372663129.0, "reward": 0.037109375, "reward_std": 0.05919293686747551, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 404 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.83984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2001.0, "completions/mean_length": 1898.0078125, "completions/mean_terminated_length": 1111.46337890625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.1382606469232739, "frac_reward_zero_std": 0.78125, "grad_norm": 13.62090865913346, "learning_rate": 9.960710637796617e-07, "loss": 0.0068, "num_tokens": 373711373.0, "reward": 0.0546875, "reward_std": 0.07685433328151703, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 405 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.873046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2034.0, "completions/mean_length": 1933.3359375, "completions/mean_terminated_length": 1144.800048828125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.13860203123666467, "frac_reward_zero_std": 0.84375, "grad_norm": 10.011669243268397, "learning_rate": 9.960000587882396e-07, "loss": 0.0104, "num_tokens": 374775161.0, "reward": 0.072265625, "reward_std": 0.05452118441462517, "rewards/accuracy_reward/mean": 0.072265625, "rewards/accuracy_reward/std": 0.2591804563999176, "step": 406 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2024.0, "completions/mean_length": 1956.77734375, "completions/mean_terminated_length": 1319.9375, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.1389434155500555, "frac_reward_zero_std": 0.875, "grad_norm": 12.79825251627794, "learning_rate": 9.95928420781309e-07, "loss": 0.0052, "num_tokens": 375863959.0, "reward": 0.0546875, "reward_std": 0.04517117142677307, "rewards/accuracy_reward/mean": 0.05645161122083664, "rewards/accuracy_reward/std": 0.23102475702762604, "step": 407 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.90234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2029.0, "completions/mean_length": 1967.76171875, "completions/mean_terminated_length": 1227.4599609375, "completions/min_length": 7.0, "completions/min_terminated_length": 7.0, "epoch": 0.13928479986344627, "frac_reward_zero_std": 0.96875, "grad_norm": 7.133796257019633, "learning_rate": 9.95856149860546e-07, "loss": 0.0048, "num_tokens": 376958429.0, "reward": 0.001953125, "reward_std": 0.0078125, "rewards/accuracy_reward/mean": 0.001953125, "rewards/accuracy_reward/std": 0.04419417306780815, "step": 408 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.931640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2028.0, "completions/mean_length": 1988.0859375, "completions/mean_terminated_length": 1171.5428466796875, "completions/min_length": 179.0, "completions/min_terminated_length": 179.0, "epoch": 0.1396261841768371, "frac_reward_zero_std": 0.90625, "grad_norm": 6.549601892863644, "learning_rate": 9.957832461285267e-07, "loss": 0.0145, "num_tokens": 378053097.0, "reward": 0.013671875, "reward_std": 0.03300705552101135, "rewards/accuracy_reward/mean": 0.013671875, "rewards/accuracy_reward/std": 0.1162383034825325, "step": 409 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.85546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1991.0, "completions/mean_length": 1927.81640625, "completions/mean_terminated_length": 1216.45947265625, "completions/min_length": 35.0, "completions/min_terminated_length": 35.0, "epoch": 0.13996756849022787, "frac_reward_zero_std": 0.84375, "grad_norm": 14.15011254657329, "learning_rate": 9.957097096887246e-07, "loss": 0.0019, "num_tokens": 379130923.0, "reward": 0.048828125, "reward_std": 0.062062256038188934, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 410 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.923828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1977.0, "completions/mean_length": 1971.41015625, "completions/mean_terminated_length": 1042.5128173828125, "completions/min_length": 279.0, "completions/min_terminated_length": 279.0, "epoch": 0.1403089528036187, "frac_reward_zero_std": 0.90625, "grad_norm": 4.760457766557874, "learning_rate": 9.95635540645511e-07, "loss": 0.0227, "num_tokens": 380217165.0, "reward": 0.041015625, "reward_std": 0.04777328670024872, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 411 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.900390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2040.0, "completions/mean_length": 1940.763671875, "completions/mean_terminated_length": 971.431396484375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.14065033711700947, "frac_reward_zero_std": 0.90625, "grad_norm": 6.7369027415487865, "learning_rate": 9.95560739104155e-07, "loss": 0.0107, "num_tokens": 381288068.0, "reward": 0.01171875, "reward_std": 0.031083684414625168, "rewards/accuracy_reward/mean": 0.012096773833036423, "rewards/accuracy_reward/std": 0.10942844301462173, "step": 412 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.81640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1889.0, "completions/mean_length": 1843.568359375, "completions/mean_terminated_length": 934.5, "completions/min_length": 13.0, "completions/min_terminated_length": 13.0, "epoch": 0.14099172143040029, "frac_reward_zero_std": 0.875, "grad_norm": 13.68441835155953, "learning_rate": 9.95485305170824e-07, "loss": 0.018, "num_tokens": 382300823.0, "reward": 0.08984375, "reward_std": 0.04742163419723511, "rewards/accuracy_reward/mean": 0.08984375, "rewards/accuracy_reward/std": 0.2862374484539032, "step": 413 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.87890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2020.0, "completions/mean_length": 1896.9375, "completions/mean_terminated_length": 800.51611328125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.14133310574379107, "frac_reward_zero_std": 0.78125, "grad_norm": 27.493703456695176, "learning_rate": 9.95409238952583e-07, "loss": 0.0277, "num_tokens": 383359575.0, "reward": 0.03515625, "reward_std": 0.07779236882925034, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 414 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.91015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1971.0, "completions/mean_length": 1961.908203125, "completions/mean_terminated_length": 1089.7608642578125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.14167449005718188, "frac_reward_zero_std": 0.875, "grad_norm": 15.649384548835311, "learning_rate": 9.953325405573935e-07, "loss": 0.014, "num_tokens": 384438072.0, "reward": 0.017578125, "reward_std": 0.04219770431518555, "rewards/accuracy_reward/mean": 0.017578125, "rewards/accuracy_reward/std": 0.13154059648513794, "step": 415 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.958984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1900.0, "completions/mean_length": 1992.30859375, "completions/mean_terminated_length": 690.1904907226562, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.14201587437057267, "frac_reward_zero_std": 0.875, "grad_norm": 6.0540331220285415, "learning_rate": 9.952552100941155e-07, "loss": 0.0203, "num_tokens": 385535862.0, "reward": 0.015625, "reward_std": 0.04081955552101135, "rewards/accuracy_reward/mean": 0.015625, "rewards/accuracy_reward/std": 0.12414088100194931, "step": 416 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.84765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1983.0, "completions/mean_length": 1907.806640625, "completions/mean_terminated_length": 1127.7564697265625, "completions/min_length": 11.0, "completions/min_terminated_length": 11.0, "epoch": 0.14235725868396348, "frac_reward_zero_std": 0.78125, "grad_norm": 10.839195975122427, "learning_rate": 9.951772476725047e-07, "loss": 0.0119, "num_tokens": 386584483.0, "reward": 0.068359375, "reward_std": 0.07669496536254883, "rewards/accuracy_reward/mean": 0.068359375, "rewards/accuracy_reward/std": 0.25260838866233826, "step": 417 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.861328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2025.0, "completions/mean_length": 1930.322265625, "completions/mean_terminated_length": 1199.394287109375, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.14269864299735427, "frac_reward_zero_std": 0.9375, "grad_norm": 2.138767764865018, "learning_rate": 9.950986534032149e-07, "loss": 0.0032, "num_tokens": 387648616.0, "reward": 0.01171875, "reward_std": 0.024649331346154213, "rewards/accuracy_reward/mean": 0.01171875, "rewards/accuracy_reward/std": 0.10772226005792618, "step": 418 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2016.0, "completions/mean_length": 1856.57421875, "completions/mean_terminated_length": 934.25, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.14304002731074508, "frac_reward_zero_std": 0.875, "grad_norm": 5.200950273517261, "learning_rate": 9.950194273977964e-07, "loss": 0.014, "num_tokens": 388671246.0, "reward": 0.0703125, "reward_std": 0.060571081936359406, "rewards/accuracy_reward/mean": 0.0703125, "rewards/accuracy_reward/std": 0.25592297315597534, "step": 419 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.939453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2011.0, "completions/mean_length": 1977.453125, "completions/mean_terminated_length": 882.8386840820312, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.14338141162413587, "frac_reward_zero_std": 0.96875, "grad_norm": 7.384092349387717, "learning_rate": 9.949395697686958e-07, "loss": 0.0013, "num_tokens": 389768150.0, "reward": 0.0078125, "reward_std": 0.013975424692034721, "rewards/accuracy_reward/mean": 0.008064515888690948, "rewards/accuracy_reward/std": 0.0895301103591919, "step": 420 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.91015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2032.0, "completions/mean_length": 1957.583984375, "completions/mean_terminated_length": 1041.6304931640625, "completions/min_length": 70.0, "completions/min_terminated_length": 70.0, "epoch": 0.14372279593752668, "frac_reward_zero_std": 0.875, "grad_norm": 8.953814546850847, "learning_rate": 9.948590806292565e-07, "loss": 0.0167, "num_tokens": 390855233.0, "reward": 0.041015625, "reward_std": 0.051707521080970764, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 421 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.892578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2035.0, "completions/mean_length": 1937.830078125, "completions/mean_terminated_length": 1022.4181518554688, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.14406418025091747, "frac_reward_zero_std": 0.78125, "grad_norm": 65.62528060669477, "learning_rate": 9.947779600937181e-07, "loss": 0.0053, "num_tokens": 391925498.0, "reward": 0.025390625, "reward_std": 0.06288585811853409, "rewards/accuracy_reward/mean": 0.025390625, "rewards/accuracy_reward/std": 0.15746226906776428, "step": 422 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.888671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1990.0, "completions/mean_length": 1934.337890625, "completions/mean_terminated_length": 1027.03515625, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.14440556456430828, "frac_reward_zero_std": 0.8125, "grad_norm": 13.415628095915988, "learning_rate": 9.946962082772163e-07, "loss": 0.0087, "num_tokens": 392993815.0, "reward": 0.03125, "reward_std": 0.06563520431518555, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 423 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1907.03125, "completions/mean_terminated_length": 1098.3157958984375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.14474694887769907, "frac_reward_zero_std": 0.8125, "grad_norm": 14.328699104696629, "learning_rate": 9.946138252957827e-07, "loss": 0.0024, "num_tokens": 394046343.0, "reward": 0.04296875, "reward_std": 0.06546889245510101, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 424 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.923828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1941.0, "completions/mean_length": 1972.025390625, "completions/mean_terminated_length": 1051.974365234375, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.14508833319108988, "frac_reward_zero_std": 0.875, "grad_norm": 2.0807841019737863, "learning_rate": 9.945308112663455e-07, "loss": 0.0124, "num_tokens": 395140260.0, "reward": 0.013671875, "reward_std": 0.03889618441462517, "rewards/accuracy_reward/mean": 0.013671875, "rewards/accuracy_reward/std": 0.1162383034825325, "step": 425 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.82421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2000.0, "completions/mean_length": 1833.947265625, "completions/mean_terminated_length": 830.2777709960938, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.14542971750448067, "frac_reward_zero_std": 0.875, "grad_norm": 7.121932787306299, "learning_rate": 9.94447166306727e-07, "loss": 0.0144, "num_tokens": 396154825.0, "reward": 0.03125, "reward_std": 0.04670868441462517, "rewards/accuracy_reward/mean": 0.032258063554763794, "rewards/accuracy_reward/std": 0.17686307430267334, "step": 426 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.880859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1966.0, "completions/mean_length": 1910.654296875, "completions/mean_terminated_length": 895.1966552734375, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.14577110181787148, "frac_reward_zero_std": 0.8125, "grad_norm": 13.671881201259376, "learning_rate": 9.94362890535647e-07, "loss": 0.0206, "num_tokens": 397211640.0, "reward": 0.03125, "reward_std": 0.06079617142677307, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 427 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.92578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1898.0, "completions/mean_length": 1991.23046875, "completions/mean_terminated_length": 1283.105224609375, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.14611248613126226, "frac_reward_zero_std": 0.84375, "grad_norm": 7.131014829230976, "learning_rate": 9.942779840727185e-07, "loss": 0.0124, "num_tokens": 398301950.0, "reward": 0.02734375, "reward_std": 0.05342378467321396, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 428 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.845703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2044.0, "completions/mean_length": 1879.447265625, "completions/mean_terminated_length": 956.088623046875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.14645387044465308, "frac_reward_zero_std": 0.9375, "grad_norm": 13.492673337281245, "learning_rate": 9.941924470384515e-07, "loss": 0.0021, "num_tokens": 399343411.0, "reward": 0.005859375, "reward_std": 0.01848640665411949, "rewards/accuracy_reward/mean": 0.005859375, "rewards/accuracy_reward/std": 0.07639661431312561, "step": 429 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1789.0, "completions/mean_length": 1891.458984375, "completions/mean_terminated_length": 796.40625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.14679525475804386, "frac_reward_zero_std": 0.78125, "grad_norm": 6.923054584476633, "learning_rate": 9.941062795542496e-07, "loss": 0.0041, "num_tokens": 400398430.0, "reward": 0.033203125, "reward_std": 0.07470491528511047, "rewards/accuracy_reward/mean": 0.033203125, "rewards/accuracy_reward/std": 0.17934183776378632, "step": 430 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.78125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2044.0, "completions/mean_length": 1856.728515625, "completions/mean_terminated_length": 1173.6160888671875, "completions/min_length": 9.0, "completions/min_terminated_length": 9.0, "epoch": 0.14713663907143468, "frac_reward_zero_std": 0.6875, "grad_norm": 31.427259022046577, "learning_rate": 9.94019481742412e-07, "loss": 0.0281, "num_tokens": 401431187.0, "reward": 0.119140625, "reward_std": 0.11819230765104294, "rewards/accuracy_reward/mean": 0.119140625, "rewards/accuracy_reward/std": 0.32427072525024414, "step": 431 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2030.0, "completions/mean_length": 1783.14453125, "completions/mean_terminated_length": 792.8518676757812, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.14747802338482546, "frac_reward_zero_std": 0.875, "grad_norm": 17.78873551003455, "learning_rate": 9.939320537261326e-07, "loss": 0.0066, "num_tokens": 402429773.0, "reward": 0.033203125, "reward_std": 0.053357094526290894, "rewards/accuracy_reward/mean": 0.03427419438958168, "rewards/accuracy_reward/std": 0.18211627006530762, "step": 432 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1948.0, "completions/mean_length": 1895.732421875, "completions/mean_terminated_length": 965.2083129882812, "completions/min_length": 8.0, "completions/min_terminated_length": 8.0, "epoch": 0.14781940769821628, "frac_reward_zero_std": 0.8125, "grad_norm": 15.93705640359269, "learning_rate": 9.93843995629499e-07, "loss": 0.0125, "num_tokens": 403483076.0, "reward": 0.08203125, "reward_std": 0.0645298957824707, "rewards/accuracy_reward/mean": 0.08203125, "rewards/accuracy_reward/std": 0.2746807038784027, "step": 433 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.85546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1976.0, "completions/mean_length": 1876.96484375, "completions/mean_terminated_length": 864.6216430664062, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.14816079201160706, "frac_reward_zero_std": 0.84375, "grad_norm": 10.436705781921948, "learning_rate": 9.937553075774938e-07, "loss": 0.0026, "num_tokens": 404521042.0, "reward": 0.013671875, "reward_std": 0.043847277760505676, "rewards/accuracy_reward/mean": 0.013671875, "rewards/accuracy_reward/std": 0.1162383034825325, "step": 434 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.927734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1899.0, "completions/mean_length": 1969.419921875, "completions/mean_terminated_length": 960.6216430664062, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.14850217632499788, "frac_reward_zero_std": 0.90625, "grad_norm": 2.108804417409502, "learning_rate": 9.936659896959935e-07, "loss": 0.0077, "num_tokens": 405608729.0, "reward": 0.0078125, "reward_std": 0.02629890665411949, "rewards/accuracy_reward/mean": 0.0078125, "rewards/accuracy_reward/std": 0.08812850713729858, "step": 435 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.771484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1989.0, "completions/mean_length": 1822.34765625, "completions/mean_terminated_length": 1060.5299072265625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.14884356063838866, "frac_reward_zero_std": 0.75, "grad_norm": 7.426146056782713, "learning_rate": 9.935760421117686e-07, "loss": 0.0366, "num_tokens": 406619403.0, "reward": 0.09375, "reward_std": 0.09594857692718506, "rewards/accuracy_reward/mean": 0.09375, "rewards/accuracy_reward/std": 0.29176566004753113, "step": 436 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.78125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1823.0, "completions/mean_length": 1780.779296875, "completions/mean_terminated_length": 826.419677734375, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.14918494495177947, "frac_reward_zero_std": 0.75, "grad_norm": 6.440456789409149, "learning_rate": 9.93485464952483e-07, "loss": 0.035, "num_tokens": 407613098.0, "reward": 0.03515625, "reward_std": 0.08157937228679657, "rewards/accuracy_reward/mean": 0.03629032149910927, "rewards/accuracy_reward/std": 0.1872003972530365, "step": 437 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.810546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1939.0, "completions/mean_length": 1848.240234375, "completions/mean_terminated_length": 994.1030883789062, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.14952632926517026, "frac_reward_zero_std": 0.8125, "grad_norm": 11.368768573442622, "learning_rate": 9.933942583466944e-07, "loss": 0.014, "num_tokens": 408639301.0, "reward": 0.033203125, "reward_std": 0.06271953880786896, "rewards/accuracy_reward/mean": 0.033203125, "rewards/accuracy_reward/std": 0.17934183776378632, "step": 438 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.85546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1987.0, "completions/mean_length": 1885.396484375, "completions/mean_terminated_length": 922.95947265625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.14986771357856107, "frac_reward_zero_std": 0.90625, "grad_norm": 4.837798002478943, "learning_rate": 9.93302422423854e-07, "loss": 0.0033, "num_tokens": 409675728.0, "reward": 0.0078125, "reward_std": 0.02629890665411949, "rewards/accuracy_reward/mean": 0.0078125, "rewards/accuracy_reward/std": 0.08812850713729858, "step": 439 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.810546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1977.0, "completions/mean_length": 1858.60546875, "completions/mean_terminated_length": 1048.3092041015625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.15020909789195186, "frac_reward_zero_std": 0.71875, "grad_norm": 25.826564229933172, "learning_rate": 9.93209957314306e-07, "loss": 0.0324, "num_tokens": 410708342.0, "reward": 0.0390625, "reward_std": 0.09506931155920029, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 440 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.78515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2007.0, "completions/mean_length": 1775.900390625, "completions/mean_terminated_length": 781.5, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.15055048220534267, "frac_reward_zero_std": 0.78125, "grad_norm": 11.985461888331729, "learning_rate": 9.93116863149288e-07, "loss": 0.0127, "num_tokens": 411685731.0, "reward": 0.052734375, "reward_std": 0.08121949434280396, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 441 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.876953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1996.0, "completions/mean_length": 1922.626953125, "completions/mean_terminated_length": 1029.857177734375, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.15089186651873346, "frac_reward_zero_std": 0.8125, "grad_norm": 11.83284768215417, "learning_rate": 9.9302314006093e-07, "loss": 0.0123, "num_tokens": 412748388.0, "reward": 0.046875, "reward_std": 0.06233368441462517, "rewards/accuracy_reward/mean": 0.04838709533214569, "rewards/accuracy_reward/std": 0.21479946374893188, "step": 442 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.791015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2020.0, "completions/mean_length": 1812.51953125, "completions/mean_terminated_length": 921.5794067382812, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.15123325083212427, "frac_reward_zero_std": 0.75, "grad_norm": 6.644910885819239, "learning_rate": 9.929287881822545e-07, "loss": 0.0057, "num_tokens": 413758702.0, "reward": 0.048828125, "reward_std": 0.08149883151054382, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 443 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.791015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1976.0, "completions/mean_length": 1829.255859375, "completions/mean_terminated_length": 1001.2990112304688, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.15157463514551506, "frac_reward_zero_std": 0.78125, "grad_norm": 12.753194935359298, "learning_rate": 9.92833807647177e-07, "loss": 0.0228, "num_tokens": 414773489.0, "reward": 0.056640625, "reward_std": 0.07729348540306091, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 444 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.876953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1908.0, "completions/mean_length": 1888.41015625, "completions/mean_terminated_length": 751.0159301757812, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.15191601945890587, "frac_reward_zero_std": 0.875, "grad_norm": 1.1789599203284744, "learning_rate": 9.927381985905051e-07, "loss": 0.0108, "num_tokens": 415817363.0, "reward": 0.017578125, "reward_std": 0.04125870764255524, "rewards/accuracy_reward/mean": 0.018145160749554634, "rewards/accuracy_reward/std": 0.1336110383272171, "step": 445 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.86328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2007.0, "completions/mean_length": 1898.666015625, "completions/mean_terminated_length": 955.7285766601562, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.15225740377229666, "frac_reward_zero_std": 0.90625, "grad_norm": 1.5471834044962816, "learning_rate": 9.92641961147938e-07, "loss": 0.0094, "num_tokens": 416864504.0, "reward": 0.013671875, "reward_std": 0.032461829483509064, "rewards/accuracy_reward/mean": 0.013671875, "rewards/accuracy_reward/std": 0.1162383034825325, "step": 446 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.87109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2023.0, "completions/mean_length": 1891.697265625, "completions/mean_terminated_length": 835.4697265625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.15259878808568747, "frac_reward_zero_std": 0.9375, "grad_norm": 1.1969890488326105, "learning_rate": 9.925450954560676e-07, "loss": 0.0117, "num_tokens": 417904909.0, "reward": 0.0078125, "reward_std": 0.020409777760505676, "rewards/accuracy_reward/mean": 0.0078125, "rewards/accuracy_reward/std": 0.08812850713729858, "step": 447 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.87109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1983.0, "completions/mean_length": 1923.06640625, "completions/mean_terminated_length": 1078.8182373046875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.15294017239907826, "frac_reward_zero_std": 0.84375, "grad_norm": 19.477288233251414, "learning_rate": 9.924476016523765e-07, "loss": 0.0101, "num_tokens": 418967311.0, "reward": 0.03125, "reward_std": 0.059520021080970764, "rewards/accuracy_reward/mean": 0.032258063554763794, "rewards/accuracy_reward/std": 0.17686307430267334, "step": 448 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.775390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2014.0, "completions/mean_length": 1826.66015625, "completions/mean_terminated_length": 1062.5565185546875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.15328155671246907, "frac_reward_zero_std": 0.875, "grad_norm": 9.476519750183675, "learning_rate": 9.9234947987524e-07, "loss": 0.0077, "num_tokens": 419973329.0, "reward": 0.025390625, "reward_std": 0.048086829483509064, "rewards/accuracy_reward/mean": 0.025390625, "rewards/accuracy_reward/std": 0.15746226906776428, "step": 449 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.798828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2005.0, "completions/mean_length": 1784.16015625, "completions/mean_terminated_length": 736.4854736328125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.15362294102585985, "frac_reward_zero_std": 0.75, "grad_norm": 26.286316535343488, "learning_rate": 9.922507302639234e-07, "loss": 0.0392, "num_tokens": 420961635.0, "reward": 0.083984375, "reward_std": 0.11329546570777893, "rewards/accuracy_reward/mean": 0.083984375, "rewards/accuracy_reward/std": 0.2776356339454651, "step": 450 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.84765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1975.0, "completions/mean_length": 1889.138671875, "completions/mean_terminated_length": 1005.2179565429688, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.15396432533925067, "frac_reward_zero_std": 0.90625, "grad_norm": 24.811244659039755, "learning_rate": 9.921513529585842e-07, "loss": 0.0072, "num_tokens": 422010362.0, "reward": 0.0390625, "reward_std": 0.031083684414625168, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 451 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.80859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2028.0, "completions/mean_length": 1799.24609375, "completions/mean_terminated_length": 748.89794921875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.15430570965264145, "frac_reward_zero_std": 0.875, "grad_norm": 8.11324895709627, "learning_rate": 9.920513481002698e-07, "loss": 0.0045, "num_tokens": 423007864.0, "reward": 0.013671875, "reward_std": 0.03741292655467987, "rewards/accuracy_reward/mean": 0.013671875, "rewards/accuracy_reward/std": 0.1162383034825325, "step": 452 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.75390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1925.0, "completions/mean_length": 1718.63671875, "completions/mean_terminated_length": 710.1032104492188, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.15464709396603227, "frac_reward_zero_std": 0.6875, "grad_norm": 21.909204175468957, "learning_rate": 9.919507158309192e-07, "loss": 0.0259, "num_tokens": 423973070.0, "reward": 0.083984375, "reward_std": 0.1175336092710495, "rewards/accuracy_reward/mean": 0.083984375, "rewards/accuracy_reward/std": 0.2776356339454651, "step": 453 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.775390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2041.0, "completions/mean_length": 1754.193359375, "completions/mean_terminated_length": 739.9216918945312, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.15498847827942305, "frac_reward_zero_std": 0.8125, "grad_norm": 16.570547002225478, "learning_rate": 9.918494562933614e-07, "loss": -0.0001, "num_tokens": 424944881.0, "reward": 0.0234375, "reward_std": 0.05920084938406944, "rewards/accuracy_reward/mean": 0.024193547666072845, "rewards/accuracy_reward/std": 0.15380479395389557, "step": 454 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1987.0, "completions/mean_length": 1914.55078125, "completions/mean_terminated_length": 734.0385131835938, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.15532986259281387, "frac_reward_zero_std": 0.96875, "grad_norm": 5.071561512225769, "learning_rate": 9.917475696313157e-07, "loss": 0.0003, "num_tokens": 426006523.0, "reward": 0.001953125, "reward_std": 0.0078125, "rewards/accuracy_reward/mean": 0.001953125, "rewards/accuracy_reward/std": 0.04419417306780815, "step": 455 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.830078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2035.0, "completions/mean_length": 1850.921875, "completions/mean_terminated_length": 888.712646484375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.15567124690620465, "frac_reward_zero_std": 0.8125, "grad_norm": 23.727008885026166, "learning_rate": 9.916450559893917e-07, "loss": 0.0038, "num_tokens": 427039507.0, "reward": 0.05078125, "reward_std": 0.0645298883318901, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 456 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.779296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1986.0, "completions/mean_length": 1777.681640625, "completions/mean_terminated_length": 823.1947021484375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.15601263121959547, "frac_reward_zero_std": 0.84375, "grad_norm": 42.17401086487023, "learning_rate": 9.915419155130886e-07, "loss": -0.0022, "num_tokens": 428020128.0, "reward": 0.0234375, "reward_std": 0.05479401722550392, "rewards/accuracy_reward/mean": 0.0234375, "rewards/accuracy_reward/std": 0.15143637359142303, "step": 457 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.78125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2044.0, "completions/mean_length": 1796.60546875, "completions/mean_terminated_length": 898.7678833007812, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.15635401553298625, "frac_reward_zero_std": 0.8125, "grad_norm": 10.485090063427458, "learning_rate": 9.914381483487957e-07, "loss": 0.009, "num_tokens": 429017670.0, "reward": 0.033203125, "reward_std": 0.05688370764255524, "rewards/accuracy_reward/mean": 0.033203125, "rewards/accuracy_reward/std": 0.17934183776378632, "step": 458 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.787109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1989.0, "completions/mean_length": 1834.505859375, "completions/mean_terminated_length": 1045.1650390625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.15669539984637706, "frac_reward_zero_std": 0.84375, "grad_norm": 2.2693779814864006, "learning_rate": 9.913337546437912e-07, "loss": 0.0114, "num_tokens": 430032249.0, "reward": 0.044921875, "reward_std": 0.060957908630371094, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 459 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.896484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1961.0, "completions/mean_length": 1922.154296875, "completions/mean_terminated_length": 832.2830200195312, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.15703678415976785, "frac_reward_zero_std": 0.9375, "grad_norm": 5.220053865493498, "learning_rate": 9.91228734546243e-07, "loss": 0.0025, "num_tokens": 431096280.0, "reward": 0.00390625, "reward_std": 0.015625, "rewards/accuracy_reward/mean": 0.00390625, "rewards/accuracy_reward/std": 0.06243881583213806, "step": 460 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.888671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2010.0, "completions/mean_length": 1930.361328125, "completions/mean_terminated_length": 992.24560546875, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.15737816847315866, "frac_reward_zero_std": 0.84375, "grad_norm": 42.12738165132376, "learning_rate": 9.911230882052082e-07, "loss": 0.0123, "num_tokens": 432162481.0, "reward": 0.04296875, "reward_std": 0.0580955371260643, "rewards/accuracy_reward/mean": 0.04435483738780022, "rewards/accuracy_reward/std": 0.2060900777578354, "step": 461 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.80859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1960.0, "completions/mean_length": 1834.3125, "completions/mean_terminated_length": 931.591796875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.15771955278654945, "frac_reward_zero_std": 0.8125, "grad_norm": 17.261598475462286, "learning_rate": 9.91016815770632e-07, "loss": 0.0098, "num_tokens": 433178289.0, "reward": 0.046875, "reward_std": 0.0737205371260643, "rewards/accuracy_reward/mean": 0.046875, "rewards/accuracy_reward/std": 0.21157780289649963, "step": 462 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.818359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1863.0, "completions/mean_length": 1850.97265625, "completions/mean_terminated_length": 963.8709716796875, "completions/min_length": 9.0, "completions/min_terminated_length": 9.0, "epoch": 0.15806093709994026, "frac_reward_zero_std": 0.90625, "grad_norm": 22.928070970916895, "learning_rate": 9.90909917393349e-07, "loss": 0.0056, "num_tokens": 434214371.0, "reward": 0.03515625, "reward_std": 0.032021719962358475, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 463 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1922.0, "completions/mean_length": 1807.453125, "completions/mean_terminated_length": 765.0833740234375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.15840232141333105, "frac_reward_zero_std": 0.84375, "grad_norm": 7.9311594819741185, "learning_rate": 9.908023932250816e-07, "loss": 0.0131, "num_tokens": 435213403.0, "reward": 0.05859375, "reward_std": 0.06327171623706818, "rewards/accuracy_reward/mean": 0.05859375, "rewards/accuracy_reward/std": 0.23509246110916138, "step": 464 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.85546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2009.0, "completions/mean_length": 1895.07421875, "completions/mean_terminated_length": 989.9189453125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.15874370572672186, "frac_reward_zero_std": 0.90625, "grad_norm": 4.263640298655056, "learning_rate": 9.906942434184411e-07, "loss": 0.0093, "num_tokens": 436257777.0, "reward": 0.041015625, "reward_std": 0.04561128467321396, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 465 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.861328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2034.0, "completions/mean_length": 1898.78125, "completions/mean_terminated_length": 971.9436645507812, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.15908509004011265, "frac_reward_zero_std": 0.78125, "grad_norm": 60.238725047809325, "learning_rate": 9.90585468126926e-07, "loss": 0.0249, "num_tokens": 437305569.0, "reward": 0.037109375, "reward_std": 0.07669496536254883, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 466 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.705078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2048.0, "completions/mean_length": 1712.150390625, "completions/mean_terminated_length": 909.2251586914062, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.15942647435350346, "frac_reward_zero_std": 0.8125, "grad_norm": 5.357874352588069, "learning_rate": 9.904760675049233e-07, "loss": 0.0063, "num_tokens": 438259646.0, "reward": 0.02734375, "reward_std": 0.06492365896701813, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 467 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.86328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2003.0, "completions/mean_length": 1872.396484375, "completions/mean_terminated_length": 763.585693359375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.15976785866689425, "frac_reward_zero_std": 0.96875, "grad_norm": 0.07308143364605568, "learning_rate": 9.903660417077069e-07, "loss": -0.0023, "num_tokens": 439288489.0, "reward": 0.009765625, "reward_std": 0.014959799125790596, "rewards/accuracy_reward/mean": 0.009765625, "rewards/accuracy_reward/std": 0.09843364357948303, "step": 468 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.763671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2015.0, "completions/mean_length": 1787.587890625, "completions/mean_terminated_length": 946.0908813476562, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.16010924298028506, "frac_reward_zero_std": 0.90625, "grad_norm": 9.696046531817968, "learning_rate": 9.90255390891438e-07, "loss": 0.0087, "num_tokens": 440278854.0, "reward": 0.025390625, "reward_std": 0.03411140665411949, "rewards/accuracy_reward/mean": 0.026209676638245583, "rewards/accuracy_reward/std": 0.1599196344614029, "step": 469 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1862.0, "completions/mean_length": 1800.759765625, "completions/mean_terminated_length": 830.8173217773438, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.16045062729367585, "frac_reward_zero_std": 0.8125, "grad_norm": 11.917488953327458, "learning_rate": 9.90144115213166e-07, "loss": 0.0056, "num_tokens": 441281563.0, "reward": 0.037109375, "reward_std": 0.06860867142677307, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 470 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.751953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2032.0, "completions/mean_length": 1766.3671875, "completions/mean_terminated_length": 912.5984497070312, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.16079201160706666, "frac_reward_zero_std": 0.8125, "grad_norm": 17.88570821411593, "learning_rate": 9.900322148308256e-07, "loss": 0.0279, "num_tokens": 442255527.0, "reward": 0.048828125, "reward_std": 0.07874394953250885, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 471 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.68359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2008.0, "completions/mean_length": 1642.8671875, "completions/mean_terminated_length": 767.5802612304688, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.16113339592045745, "frac_reward_zero_std": 0.75, "grad_norm": 14.891304098682305, "learning_rate": 9.899196899032393e-07, "loss": 0.0146, "num_tokens": 443162387.0, "reward": 0.0625, "reward_std": 0.10067769140005112, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.2422981858253479, "step": 472 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1897.0, "completions/mean_length": 1762.927734375, "completions/mean_terminated_length": 870.9273681640625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.16147478023384826, "frac_reward_zero_std": 0.78125, "grad_norm": 10.102665470538868, "learning_rate": 9.898065405901156e-07, "loss": 0.0242, "num_tokens": 444143870.0, "reward": 0.068359375, "reward_std": 0.09078246355056763, "rewards/accuracy_reward/mean": 0.068359375, "rewards/accuracy_reward/std": 0.25260838866233826, "step": 473 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.771484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2044.0, "completions/mean_length": 1775.95703125, "completions/mean_terminated_length": 857.5214233398438, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.16181616454723904, "frac_reward_zero_std": 0.84375, "grad_norm": 6.937774072950919, "learning_rate": 9.896927670520495e-07, "loss": 0.032, "num_tokens": 445121864.0, "reward": 0.072265625, "reward_std": 0.0683835819363594, "rewards/accuracy_reward/mean": 0.072265625, "rewards/accuracy_reward/std": 0.2591804563999176, "step": 474 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.802734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2001.0, "completions/mean_length": 1794.95703125, "completions/mean_terminated_length": 765.2474975585938, "completions/min_length": 8.0, "completions/min_terminated_length": 8.0, "epoch": 0.16215754886062986, "frac_reward_zero_std": 0.875, "grad_norm": 4.331967286884289, "learning_rate": 9.895783694505212e-07, "loss": 0.0085, "num_tokens": 446117474.0, "reward": 0.01953125, "reward_std": 0.04505910724401474, "rewards/accuracy_reward/mean": 0.01953125, "rewards/accuracy_reward/std": 0.1385180652141571, "step": 475 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1955.0, "completions/mean_length": 1722.34765625, "completions/mean_terminated_length": 857.0428466796875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.16249893317402064, "frac_reward_zero_std": 0.8125, "grad_norm": 5.351624507032269, "learning_rate": 9.894633479478974e-07, "loss": 0.0224, "num_tokens": 447076516.0, "reward": 0.05078125, "reward_std": 0.07340699434280396, "rewards/accuracy_reward/mean": 0.052419353276491165, "rewards/accuracy_reward/std": 0.22309619188308716, "step": 476 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.623046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.0, "completions/mean_length": 1603.806640625, "completions/mean_terminated_length": 869.6217651367188, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.16284031748741146, "frac_reward_zero_std": 0.8125, "grad_norm": 3.8560699741509796, "learning_rate": 9.893477027074303e-07, "loss": 0.014, "num_tokens": 447975009.0, "reward": 0.0625, "reward_std": 0.0645298883318901, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.2422981858253479, "step": 477 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2022.0, "completions/mean_length": 1776.279296875, "completions/mean_terminated_length": 994.0530395507812, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.16318170180080224, "frac_reward_zero_std": 0.8125, "grad_norm": 6.631906826141045, "learning_rate": 9.89231433893257e-07, "loss": 0.0201, "num_tokens": 448961504.0, "reward": 0.029296875, "reward_std": 0.064970001578331, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 478 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.81640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2040.0, "completions/mean_length": 1843.33203125, "completions/mean_terminated_length": 933.2127075195312, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.16352308611419306, "frac_reward_zero_std": 0.875, "grad_norm": 0.48690598958937414, "learning_rate": 9.891145416703998e-07, "loss": 0.0033, "num_tokens": 449978746.0, "reward": 0.009765625, "reward_std": 0.03411140665411949, "rewards/accuracy_reward/mean": 0.009765625, "rewards/accuracy_reward/std": 0.09843364357948303, "step": 479 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.86328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2031.0, "completions/mean_length": 1895.099609375, "completions/mean_terminated_length": 929.6428833007812, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.16386447042758384, "frac_reward_zero_std": 0.90625, "grad_norm": 0.6989406952855886, "learning_rate": 9.889970262047658e-07, "loss": 0.0099, "num_tokens": 451026973.0, "reward": 0.013671875, "reward_std": 0.032461829483509064, "rewards/accuracy_reward/mean": 0.013671875, "rewards/accuracy_reward/std": 0.1162383034825325, "step": 480 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.767578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2039.0, "completions/mean_length": 1775.75390625, "completions/mean_terminated_length": 876.655517578125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.16420585474097465, "frac_reward_zero_std": 0.8125, "grad_norm": 19.866435026265687, "learning_rate": 9.888788876631467e-07, "loss": 0.0061, "num_tokens": 452010623.0, "reward": 0.033203125, "reward_std": 0.0666455551981926, "rewards/accuracy_reward/mean": 0.03427419438958168, "rewards/accuracy_reward/std": 0.18211629986763, "step": 481 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.830078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2041.0, "completions/mean_length": 1846.037109375, "completions/mean_terminated_length": 859.436767578125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.16454723905436544, "frac_reward_zero_std": 0.90625, "grad_norm": 0.24443680707577523, "learning_rate": 9.887601262132187e-07, "loss": 0.0046, "num_tokens": 453039778.0, "reward": 0.017578125, "reward_std": 0.03724660724401474, "rewards/accuracy_reward/mean": 0.017578125, "rewards/accuracy_reward/std": 0.13154059648513794, "step": 482 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.77734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2018.0, "completions/mean_length": 1794.853515625, "completions/mean_terminated_length": 911.0614013671875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.16488862336775625, "frac_reward_zero_std": 0.90625, "grad_norm": 0.6697871523979727, "learning_rate": 9.88640742023542e-07, "loss": 0.0113, "num_tokens": 454031239.0, "reward": 0.009765625, "reward_std": 0.028222277760505676, "rewards/accuracy_reward/mean": 0.009765625, "rewards/accuracy_reward/std": 0.09843364357948303, "step": 483 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.66015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1669.65234375, "completions/mean_terminated_length": 934.701171875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.16523000768114704, "frac_reward_zero_std": 0.78125, "grad_norm": 7.445078771953974, "learning_rate": 9.8852073526356e-07, "loss": 0.0118, "num_tokens": 454966309.0, "reward": 0.05859375, "reward_std": 0.0936504602432251, "rewards/accuracy_reward/mean": 0.060483869165182114, "rewards/accuracy_reward/std": 0.2386218160390854, "step": 484 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.728515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2036.0, "completions/mean_length": 1756.134765625, "completions/mean_terminated_length": 972.9281005859375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.16557139199453785, "frac_reward_zero_std": 0.8125, "grad_norm": 24.138076351078794, "learning_rate": 9.884001061036013e-07, "loss": -0.0058, "num_tokens": 455956650.0, "reward": 0.080078125, "reward_std": 0.0809411108493805, "rewards/accuracy_reward/mean": 0.080078125, "rewards/accuracy_reward/std": 0.271679550409317, "step": 485 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2045.0, "completions/mean_length": 1825.087890625, "completions/mean_terminated_length": 950.5865478515625, "completions/min_length": 8.0, "completions/min_terminated_length": 8.0, "epoch": 0.16591277630792864, "frac_reward_zero_std": 0.84375, "grad_norm": 7.1733387609781065, "learning_rate": 9.882788547148764e-07, "loss": 0.0154, "num_tokens": 456972423.0, "reward": 0.01171875, "reward_std": 0.04192390665411949, "rewards/accuracy_reward/mean": 0.012096773833036423, "rewards/accuracy_reward/std": 0.10942844301462173, "step": 486 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1923.0, "completions/mean_length": 1626.744140625, "completions/mean_terminated_length": 924.6510620117188, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.16625416062131945, "frac_reward_zero_std": 0.8125, "grad_norm": 10.91214937309718, "learning_rate": 9.881569812694795e-07, "loss": 0.0176, "num_tokens": 457882036.0, "reward": 0.048828125, "reward_std": 0.07493096590042114, "rewards/accuracy_reward/mean": 0.05040322616696358, "rewards/accuracy_reward/std": 0.21899642050266266, "step": 487 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2032.0, "completions/mean_length": 1674.9140625, "completions/mean_terminated_length": 854.125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.16659554493471024, "frac_reward_zero_std": 0.875, "grad_norm": 13.682206703302473, "learning_rate": 9.880344859403876e-07, "loss": 0.0103, "num_tokens": 458813960.0, "reward": 0.02734375, "reward_std": 0.04907120764255524, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 488 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.693359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2020.0, "completions/mean_length": 1673.44140625, "completions/mean_terminated_length": 826.5095825195312, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.16693692924810105, "frac_reward_zero_std": 0.84375, "grad_norm": 1.3641632511405746, "learning_rate": 9.879113689014606e-07, "loss": -0.0001, "num_tokens": 459755818.0, "reward": 0.041015625, "reward_std": 0.05793476477265358, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 489 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.712890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2048.0, "completions/mean_length": 1697.404296875, "completions/mean_terminated_length": 826.8775634765625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.16727831356149184, "frac_reward_zero_std": 0.65625, "grad_norm": 32.2314419948305, "learning_rate": 9.877876303274404e-07, "loss": 0.0127, "num_tokens": 460703209.0, "reward": 0.05859375, "reward_std": 0.1156454086303711, "rewards/accuracy_reward/mean": 0.05859375, "rewards/accuracy_reward/std": 0.23509246110916138, "step": 490 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.673828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2045.0, "completions/mean_length": 1649.765625, "completions/mean_terminated_length": 827.06591796875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.16761969787488265, "frac_reward_zero_std": 0.71875, "grad_norm": 16.270410936155635, "learning_rate": 9.876632703939517e-07, "loss": 0.0112, "num_tokens": 461622097.0, "reward": 0.046875, "reward_std": 0.09518137574195862, "rewards/accuracy_reward/mean": 0.046875, "rewards/accuracy_reward/std": 0.21157780289649963, "step": 491 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.771484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2003.0, "completions/mean_length": 1789.486328125, "completions/mean_terminated_length": 916.7265625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.16796108218827344, "frac_reward_zero_std": 0.75, "grad_norm": 2.639480166458279, "learning_rate": 9.875382892775e-07, "loss": 0.021, "num_tokens": 462613258.0, "reward": 0.025390625, "reward_std": 0.07438573986291885, "rewards/accuracy_reward/mean": 0.025390625, "rewards/accuracy_reward/std": 0.15746226906776428, "step": 492 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.78125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2039.0, "completions/mean_length": 1836.68359375, "completions/mean_terminated_length": 1081.982177734375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.16830246650166425, "frac_reward_zero_std": 0.78125, "grad_norm": 17.060624547414804, "learning_rate": 9.874126871554738e-07, "loss": 0.0224, "num_tokens": 463628632.0, "reward": 0.033203125, "reward_std": 0.07586899399757385, "rewards/accuracy_reward/mean": 0.033203125, "rewards/accuracy_reward/std": 0.17934183776378632, "step": 493 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.830078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1957.0, "completions/mean_length": 1818.36328125, "completions/mean_terminated_length": 696.57470703125, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.16864385081505504, "frac_reward_zero_std": 0.78125, "grad_norm": 26.319974956419962, "learning_rate": 9.872864642061419e-07, "loss": 0.0057, "num_tokens": 464628306.0, "reward": 0.025390625, "reward_std": 0.06849660724401474, "rewards/accuracy_reward/mean": 0.026209676638245583, "rewards/accuracy_reward/std": 0.1599196493625641, "step": 494 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.67578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2037.0, "completions/mean_length": 1614.453125, "completions/mean_terminated_length": 710.795166015625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.16898523512844585, "frac_reward_zero_std": 0.78125, "grad_norm": 26.292361686925275, "learning_rate": 9.87159620608655e-07, "loss": 0.0157, "num_tokens": 465536026.0, "reward": 0.060546875, "reward_std": 0.0840085819363594, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.24230584502220154, "step": 495 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.79296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2035.0, "completions/mean_length": 1807.380859375, "completions/mean_terminated_length": 885.76416015625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.16932661944183663, "frac_reward_zero_std": 0.84375, "grad_norm": 10.90641434067934, "learning_rate": 9.87032156543044e-07, "loss": 0.0224, "num_tokens": 466532221.0, "reward": 0.021484375, "reward_std": 0.054354868829250336, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 496 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.80078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1999.0, "completions/mean_length": 1827.595703125, "completions/mean_terminated_length": 941.6569213867188, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.16966800375522745, "frac_reward_zero_std": 0.8125, "grad_norm": 13.087614501863381, "learning_rate": 9.869040721902213e-07, "loss": 0.0023, "num_tokens": 467544654.0, "reward": 0.03125, "reward_std": 0.06409768760204315, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 497 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.779296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.0, "completions/mean_length": 1827.9765625, "completions/mean_terminated_length": 1051.07958984375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.17000938806861823, "frac_reward_zero_std": 0.8125, "grad_norm": 27.88031154057963, "learning_rate": 9.86775367731979e-07, "loss": -0.004, "num_tokens": 468561122.0, "reward": 0.083984375, "reward_std": 0.08094111829996109, "rewards/accuracy_reward/mean": 0.083984375, "rewards/accuracy_reward/std": 0.2776356339454651, "step": 498 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.787109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2004.0, "completions/mean_length": 1815.07421875, "completions/mean_terminated_length": 953.889892578125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.17035077238200905, "frac_reward_zero_std": 0.875, "grad_norm": 36.14980895384696, "learning_rate": 9.866460433509893e-07, "loss": 0.0106, "num_tokens": 469566328.0, "reward": 0.033203125, "reward_std": 0.054241843521595, "rewards/accuracy_reward/mean": 0.033203125, "rewards/accuracy_reward/std": 0.17934183776378632, "step": 499 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.779296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2030.0, "completions/mean_length": 1792.544921875, "completions/mean_terminated_length": 890.539794921875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.17069215669539983, "frac_reward_zero_std": 0.75, "grad_norm": 27.01887296968589, "learning_rate": 9.865160992308047e-07, "loss": 0.0179, "num_tokens": 470558703.0, "reward": 0.091796875, "reward_std": 0.09534074366092682, "rewards/accuracy_reward/mean": 0.091796875, "rewards/accuracy_reward/std": 0.289021372795105, "step": 500 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.80859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2004.0, "completions/mean_length": 1877.0546875, "completions/mean_terminated_length": 1154.89794921875, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.17103354100879065, "frac_reward_zero_std": 0.78125, "grad_norm": 59.305129030936534, "learning_rate": 9.863855355558573e-07, "loss": 0.0144, "num_tokens": 471599627.0, "reward": 0.09765625, "reward_std": 0.08499295264482498, "rewards/accuracy_reward/mean": 0.09765625, "rewards/accuracy_reward/std": 0.29713961482048035, "step": 501 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.814453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2024.0, "completions/mean_length": 1829.72265625, "completions/mean_terminated_length": 871.6000366210938, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.17137492532218146, "frac_reward_zero_std": 0.84375, "grad_norm": 42.46915295445329, "learning_rate": 9.862543525114582e-07, "loss": 0.0092, "num_tokens": 472609997.0, "reward": 0.021484375, "reward_std": 0.05193261429667473, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 502 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.77734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2038.0, "completions/mean_length": 1826.541015625, "completions/mean_terminated_length": 1053.377197265625, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.17171630963557225, "frac_reward_zero_std": 0.875, "grad_norm": 17.236420763000325, "learning_rate": 9.861225502837976e-07, "loss": 0.0014, "num_tokens": 473628754.0, "reward": 0.029296875, "reward_std": 0.04726085811853409, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 503 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2038.0, "completions/mean_length": 1765.7109375, "completions/mean_terminated_length": 843.5667114257812, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.17205769394896306, "frac_reward_zero_std": 0.84375, "grad_norm": 11.058591294531599, "learning_rate": 9.859901290599448e-07, "loss": 0.025, "num_tokens": 474611182.0, "reward": 0.048828125, "reward_std": 0.07302350550889969, "rewards/accuracy_reward/mean": 0.05040322616696358, "rewards/accuracy_reward/std": 0.21899642050266266, "step": 504 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.85546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2002.0, "completions/mean_length": 1879.12890625, "completions/mean_terminated_length": 879.5946044921875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.17239907826235384, "frac_reward_zero_std": 0.8125, "grad_norm": 12.255340365910076, "learning_rate": 9.858570890278475e-07, "loss": 0.0121, "num_tokens": 475647424.0, "reward": 0.03125, "reward_std": 0.06837663054466248, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 505 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.865234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2045.0, "completions/mean_length": 1868.5078125, "completions/mean_terminated_length": 716.115966796875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.17274046257574466, "frac_reward_zero_std": 0.875, "grad_norm": 7.772987306282009, "learning_rate": 9.857234303763317e-07, "loss": 0.0187, "num_tokens": 476682724.0, "reward": 0.037109375, "reward_std": 0.05099457502365112, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 506 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2019.0, "completions/mean_length": 1883.412109375, "completions/mean_terminated_length": 808.75, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.17308184688913544, "frac_reward_zero_std": 0.875, "grad_norm": 6.656879661902198, "learning_rate": 9.855891532951015e-07, "loss": 0.0159, "num_tokens": 477727463.0, "reward": 0.01171875, "reward_std": 0.03697281330823898, "rewards/accuracy_reward/mean": 0.01171875, "rewards/accuracy_reward/std": 0.10772226005792618, "step": 507 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2031.0, "completions/mean_length": 1884.251953125, "completions/mean_terminated_length": 883.5694580078125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.17342323120252626, "frac_reward_zero_std": 0.90625, "grad_norm": 10.459997247741349, "learning_rate": 9.854542579747383e-07, "loss": 0.0051, "num_tokens": 478774088.0, "reward": 0.0078125, "reward_std": 0.02629890665411949, "rewards/accuracy_reward/mean": 0.0078125, "rewards/accuracy_reward/std": 0.08812850713729858, "step": 508 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1943.6875, "completions/mean_terminated_length": 1095.071533203125, "completions/min_length": 9.0, "completions/min_terminated_length": 9.0, "epoch": 0.17376461551591704, "frac_reward_zero_std": 0.96875, "grad_norm": 1.482333780461527, "learning_rate": 9.853187446067019e-07, "loss": 0.0055, "num_tokens": 479856200.0, "reward": 0.001953125, "reward_std": 0.0078125, "rewards/accuracy_reward/mean": 0.001953125, "rewards/accuracy_reward/std": 0.04419417306780815, "step": 509 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1850.0, "completions/mean_length": 1923.134765625, "completions/mean_terminated_length": 906.3750610351562, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.17410599982930786, "frac_reward_zero_std": 0.875, "grad_norm": 5.602772611966746, "learning_rate": 9.85182613383328e-07, "loss": 0.0121, "num_tokens": 480914365.0, "reward": 0.013671875, "reward_std": 0.03889618441462517, "rewards/accuracy_reward/mean": 0.013671875, "rewards/accuracy_reward/std": 0.1162383034825325, "step": 510 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.876953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2030.0, "completions/mean_length": 1895.380859375, "completions/mean_terminated_length": 807.6666870117188, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.17444738414269864, "frac_reward_zero_std": 0.9375, "grad_norm": 1.3264946718671955, "learning_rate": 9.850458644978307e-07, "loss": 0.0114, "num_tokens": 481964096.0, "reward": 0.009765625, "reward_std": 0.023271184414625168, "rewards/accuracy_reward/mean": 0.009765625, "rewards/accuracy_reward/std": 0.09843364357948303, "step": 511 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.873046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2030.0, "completions/mean_length": 1900.189453125, "completions/mean_terminated_length": 883.7077026367188, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.17478876845608946, "frac_reward_zero_std": 0.96875, "grad_norm": 5.089345314160676, "learning_rate": 9.849084981442997e-07, "loss": -0.0001, "num_tokens": 483012833.0, "reward": 0.009765625, "reward_std": 0.014959799125790596, "rewards/accuracy_reward/mean": 0.009765625, "rewards/accuracy_reward/std": 0.09843364357948303, "step": 512 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2039.0, "completions/mean_length": 1918.6953125, "completions/mean_terminated_length": 1013.5625, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.17513015276948024, "frac_reward_zero_std": 0.84375, "grad_norm": 46.126518427241216, "learning_rate": 9.847705145177013e-07, "loss": 0.0067, "num_tokens": 484079541.0, "reward": 0.01953125, "reward_std": 0.051493462175130844, "rewards/accuracy_reward/mean": 0.01953125, "rewards/accuracy_reward/std": 0.1385180652141571, "step": 513 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.955078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1722.0, "completions/mean_length": 1989.291015625, "completions/mean_terminated_length": 741.0869750976562, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.17547153708287105, "frac_reward_zero_std": 0.9375, "grad_norm": 4.306278338996831, "learning_rate": 9.84631913813878e-07, "loss": 0.0023, "num_tokens": 485177050.0, "reward": 0.00390625, "reward_std": 0.015625, "rewards/accuracy_reward/mean": 0.00390625, "rewards/accuracy_reward/std": 0.06243881583213806, "step": 514 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1825.0, "completions/mean_length": 1833.232421875, "completions/mean_terminated_length": 521.2916870117188, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.17581292139626184, "frac_reward_zero_std": 0.75, "grad_norm": 17.07934141698569, "learning_rate": 9.844926962295487e-07, "loss": 0.0172, "num_tokens": 486193153.0, "reward": 0.046875, "reward_std": 0.08345641195774078, "rewards/accuracy_reward/mean": 0.046875, "rewards/accuracy_reward/std": 0.21157780289649963, "step": 515 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2031.0, "completions/mean_length": 1886.50390625, "completions/mean_terminated_length": 1063.642822265625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.17615430570965265, "frac_reward_zero_std": 0.84375, "grad_norm": 25.918379914242184, "learning_rate": 9.843528619623068e-07, "loss": -0.0068, "num_tokens": 487242851.0, "reward": 0.048828125, "reward_std": 0.055459219962358475, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 516 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.908203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1968.0, "completions/mean_length": 1945.52734375, "completions/mean_terminated_length": 931.7020874023438, "completions/min_length": 11.0, "completions/min_terminated_length": 11.0, "epoch": 0.17649569002304344, "frac_reward_zero_std": 0.90625, "grad_norm": 6.074489289766609, "learning_rate": 9.842124112106214e-07, "loss": 0.0136, "num_tokens": 488316257.0, "reward": 0.0234375, "reward_std": 0.03449726849794388, "rewards/accuracy_reward/mean": 0.0234375, "rewards/accuracy_reward/std": 0.15143637359142303, "step": 517 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.873046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1961.0, "completions/mean_length": 1909.5859375, "completions/mean_terminated_length": 958.4000244140625, "completions/min_length": 9.0, "completions/min_terminated_length": 9.0, "epoch": 0.17683707433643425, "frac_reward_zero_std": 0.90625, "grad_norm": 7.7535261999195635, "learning_rate": 9.84071344173837e-07, "loss": 0.0016, "num_tokens": 489378333.0, "reward": 0.025390625, "reward_std": 0.036420635879039764, "rewards/accuracy_reward/mean": 0.025390625, "rewards/accuracy_reward/std": 0.15746226906776428, "step": 518 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.888671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1794.0, "completions/mean_length": 1894.037109375, "completions/mean_terminated_length": 665.0350952148438, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.17717845864982504, "frac_reward_zero_std": 0.90625, "grad_norm": 7.51208727068919, "learning_rate": 9.839296610521723e-07, "loss": 0.0062, "num_tokens": 490430016.0, "reward": 0.005859375, "reward_std": 0.0234375, "rewards/accuracy_reward/mean": 0.005859375, "rewards/accuracy_reward/std": 0.07639661431312561, "step": 519 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.869140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2016.0, "completions/mean_length": 1881.80078125, "completions/mean_terminated_length": 777.9403076171875, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.17751984296321585, "frac_reward_zero_std": 0.875, "grad_norm": 9.310692694221732, "learning_rate": 9.837873620467203e-07, "loss": 0.0215, "num_tokens": 491472858.0, "reward": 0.03125, "reward_std": 0.04738742858171463, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 520 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.84765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2036.0, "completions/mean_length": 1865.65625, "completions/mean_terminated_length": 851.076904296875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.17786122727660664, "frac_reward_zero_std": 0.875, "grad_norm": 10.328715596479137, "learning_rate": 9.836444473594488e-07, "loss": 0.0129, "num_tokens": 492511274.0, "reward": 0.0234375, "reward_std": 0.04670868441462517, "rewards/accuracy_reward/mean": 0.0234375, "rewards/accuracy_reward/std": 0.15143637359142303, "step": 521 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.896484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1896.0, "completions/mean_length": 1898.76171875, "completions/mean_terminated_length": 606.3018798828125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.17820261158999745, "frac_reward_zero_std": 0.96875, "grad_norm": 12.579588330372541, "learning_rate": 9.83500917193199e-07, "loss": -0.0017, "num_tokens": 493560880.0, "reward": 0.001953125, "reward_std": 0.0078125, "rewards/accuracy_reward/mean": 0.001953125, "rewards/accuracy_reward/std": 0.04419417306780815, "step": 522 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.90625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1811.0, "completions/mean_length": 1922.841796875, "completions/mean_terminated_length": 712.9791870117188, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.17854399590338824, "frac_reward_zero_std": 0.9375, "grad_norm": 1.4255374368916374, "learning_rate": 9.833567717516856e-07, "loss": 0.0109, "num_tokens": 494619263.0, "reward": 0.0078125, "reward_std": 0.020409777760505676, "rewards/accuracy_reward/mean": 0.0078125, "rewards/accuracy_reward/std": 0.08812850713729858, "step": 523 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.857421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1941.0, "completions/mean_length": 1842.72265625, "completions/mean_terminated_length": 608.24658203125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.17888538021677905, "frac_reward_zero_std": 0.84375, "grad_norm": 5.922419274012383, "learning_rate": 9.832120112394969e-07, "loss": 0.0089, "num_tokens": 495637329.0, "reward": 0.0234375, "reward_std": 0.04738742858171463, "rewards/accuracy_reward/mean": 0.0234375, "rewards/accuracy_reward/std": 0.15143637359142303, "step": 524 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.919921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2022.0, "completions/mean_length": 1941.068359375, "completions/mean_terminated_length": 712.6585083007812, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.17922676453016984, "frac_reward_zero_std": 0.90625, "grad_norm": 9.525964094930613, "learning_rate": 9.830666358620936e-07, "loss": 0.019, "num_tokens": 496712276.0, "reward": 0.0234375, "reward_std": 0.03779878467321396, "rewards/accuracy_reward/mean": 0.0234375, "rewards/accuracy_reward/std": 0.15143637359142303, "step": 525 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2045.0, "completions/mean_length": 1917.82421875, "completions/mean_terminated_length": 857.8214721679688, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.17956814884356065, "frac_reward_zero_std": 0.875, "grad_norm": 10.077388356284844, "learning_rate": 9.829206458258097e-07, "loss": 0.0207, "num_tokens": 497772922.0, "reward": 0.0234375, "reward_std": 0.04670868441462517, "rewards/accuracy_reward/mean": 0.0234375, "rewards/accuracy_reward/std": 0.15143637359142303, "step": 526 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2039.0, "completions/mean_length": 1938.634765625, "completions/mean_terminated_length": 1225.176513671875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.17990953315695143, "frac_reward_zero_std": 0.8125, "grad_norm": 15.035390223215256, "learning_rate": 9.827740413378513e-07, "loss": 0.0119, "num_tokens": 498848239.0, "reward": 0.05078125, "reward_std": 0.06563520431518555, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 527 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.814453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1797.0, "completions/mean_length": 1808.3046875, "completions/mean_terminated_length": 756.16845703125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.18025091747034225, "frac_reward_zero_std": 0.9375, "grad_norm": 4.417740476227587, "learning_rate": 9.826268226062967e-07, "loss": 0.0041, "num_tokens": 499845467.0, "reward": 0.029296875, "reward_std": 0.01848640665411949, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 528 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.92578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1832.0, "completions/mean_length": 1923.798828125, "completions/mean_terminated_length": 374.5526428222656, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.18059230178373303, "frac_reward_zero_std": 1.0, "grad_norm": 0.0, "learning_rate": 9.82478989840096e-07, "loss": 0.0, "num_tokens": 500902916.0, "reward": 0.0, "reward_std": 0.0, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "step": 529 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.904296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1928.0, "completions/mean_length": 1925.734375, "completions/mean_terminated_length": 770.448974609375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.18093368609712385, "frac_reward_zero_std": 0.90625, "grad_norm": 7.997138510877159, "learning_rate": 9.823305432490705e-07, "loss": 0.0153, "num_tokens": 501968012.0, "reward": 0.021484375, "reward_std": 0.03741292655467987, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 530 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2016.0, "completions/mean_length": 1776.5546875, "completions/mean_terminated_length": 711.6538696289062, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.18127507041051463, "frac_reward_zero_std": 0.8125, "grad_norm": 11.314160850450367, "learning_rate": 9.821814830439133e-07, "loss": 0.0171, "num_tokens": 502958200.0, "reward": 0.037109375, "reward_std": 0.07273615896701813, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 531 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2042.0, "completions/mean_length": 1893.259765625, "completions/mean_terminated_length": 947.625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.18161645472390545, "frac_reward_zero_std": 0.84375, "grad_norm": 23.717393891089788, "learning_rate": 9.820318094361883e-07, "loss": 0.0044, "num_tokens": 504011549.0, "reward": 0.064453125, "reward_std": 0.06244116649031639, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 532 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.81640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1974.0, "completions/mean_length": 1809.103515625, "completions/mean_terminated_length": 746.7765502929688, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.18195783903729623, "frac_reward_zero_std": 0.875, "grad_norm": 7.037442383874556, "learning_rate": 9.81881522638329e-07, "loss": 0.0107, "num_tokens": 505012722.0, "reward": 0.015625, "reward_std": 0.03839729726314545, "rewards/accuracy_reward/mean": 0.015625, "rewards/accuracy_reward/std": 0.12414088100194931, "step": 533 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.884765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2012.0, "completions/mean_length": 1889.455078125, "completions/mean_terminated_length": 672.1525268554688, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.18229922335068705, "frac_reward_zero_std": 0.90625, "grad_norm": 9.081106926428404, "learning_rate": 9.817306228636411e-07, "loss": 0.0188, "num_tokens": 506049611.0, "reward": 0.021484375, "reward_std": 0.03889618441462517, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 534 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.794921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1871.0, "completions/mean_length": 1751.1015625, "completions/mean_terminated_length": 600.2666625976562, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.18264060766407783, "frac_reward_zero_std": 0.84375, "grad_norm": 10.319436215537658, "learning_rate": 9.815791103262981e-07, "loss": 0.0008, "num_tokens": 507020799.0, "reward": 0.03515625, "reward_std": 0.05931950360536575, "rewards/accuracy_reward/mean": 0.03629032149910927, "rewards/accuracy_reward/std": 0.1872003972530365, "step": 535 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.775390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2011.0, "completions/mean_length": 1775.626953125, "completions/mean_terminated_length": 835.7042846679688, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.18298199197746864, "frac_reward_zero_std": 0.84375, "grad_norm": 7.975147957281659, "learning_rate": 9.814269852413453e-07, "loss": 0.0049, "num_tokens": 508007712.0, "reward": 0.05859375, "reward_std": 0.06634815037250519, "rewards/accuracy_reward/mean": 0.060483869165182114, "rewards/accuracy_reward/std": 0.2386218160390854, "step": 536 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.818359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2025.0, "completions/mean_length": 1830.619140625, "completions/mean_terminated_length": 851.236572265625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.18332337629085943, "frac_reward_zero_std": 0.90625, "grad_norm": 2.2311419512950486, "learning_rate": 9.812742478246957e-07, "loss": 0.0009, "num_tokens": 509016813.0, "reward": 0.0234375, "reward_std": 0.036547206342220306, "rewards/accuracy_reward/mean": 0.0234375, "rewards/accuracy_reward/std": 0.15143637359142303, "step": 537 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.818359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1939.0, "completions/mean_length": 1797.908203125, "completions/mean_terminated_length": 671.1505126953125, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.18366476060425024, "frac_reward_zero_std": 0.875, "grad_norm": 5.654129346242353, "learning_rate": 9.811208982931327e-07, "loss": 0.0194, "num_tokens": 510007390.0, "reward": 0.01171875, "reward_std": 0.036034777760505676, "rewards/accuracy_reward/mean": 0.01171875, "rewards/accuracy_reward/std": 0.10772226005792618, "step": 538 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.81640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1973.0, "completions/mean_length": 1813.591796875, "completions/mean_terminated_length": 771.223388671875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.18400614491764103, "frac_reward_zero_std": 0.8125, "grad_norm": 10.112976580884771, "learning_rate": 9.809669368643075e-07, "loss": 0.0161, "num_tokens": 511012125.0, "reward": 0.033203125, "reward_std": 0.06844235956668854, "rewards/accuracy_reward/mean": 0.033203125, "rewards/accuracy_reward/std": 0.17934183776378632, "step": 539 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.806640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1793.0, "completions/mean_length": 1803.060546875, "completions/mean_terminated_length": 781.242431640625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.18434752923103184, "frac_reward_zero_std": 0.90625, "grad_norm": 4.799750780626729, "learning_rate": 9.808123637567406e-07, "loss": 0.0201, "num_tokens": 512008268.0, "reward": 0.02734375, "reward_std": 0.03697281330823898, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 540 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.787109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2021.0, "completions/mean_length": 1810.48046875, "completions/mean_terminated_length": 932.3118896484375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.18468891354442263, "frac_reward_zero_std": 0.8125, "grad_norm": 6.498139135675712, "learning_rate": 9.806571791898196e-07, "loss": 0.0112, "num_tokens": 513015698.0, "reward": 0.013671875, "reward_std": 0.04973640665411949, "rewards/accuracy_reward/mean": 0.013671875, "rewards/accuracy_reward/std": 0.1162383034825325, "step": 541 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.84375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1699.0, "completions/mean_length": 1825.5078125, "completions/mean_terminated_length": 624.0499877929688, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.18503029785781344, "frac_reward_zero_std": 0.84375, "grad_norm": 15.349433768378061, "learning_rate": 9.805013833838014e-07, "loss": 0.015, "num_tokens": 514033446.0, "reward": 0.0390625, "reward_std": 0.060571081936359406, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 542 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1994.0, "completions/mean_length": 1857.458984375, "completions/mean_terminated_length": 886.607177734375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.18537168217120423, "frac_reward_zero_std": 0.84375, "grad_norm": 8.652111917404474, "learning_rate": 9.80344976559809e-07, "loss": 0.0238, "num_tokens": 515066049.0, "reward": 0.052734375, "reward_std": 0.06397771090269089, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 543 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.81640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2031.0, "completions/mean_length": 1842.611328125, "completions/mean_terminated_length": 929.2872314453125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.18571306648459504, "frac_reward_zero_std": 0.71875, "grad_norm": 26.82701428176531, "learning_rate": 9.801879589398338e-07, "loss": 0.0113, "num_tokens": 516081850.0, "reward": 0.0546875, "reward_std": 0.09782323986291885, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 544 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.73828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2009.0, "completions/mean_length": 1738.3828125, "completions/mean_terminated_length": 864.9850463867188, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.18605445079798583, "frac_reward_zero_std": 0.65625, "grad_norm": 21.662496615970383, "learning_rate": 9.80030330746733e-07, "loss": 0.0112, "num_tokens": 517054334.0, "reward": 0.09765625, "reward_std": 0.12384376674890518, "rewards/accuracy_reward/mean": 0.09765625, "rewards/accuracy_reward/std": 0.29713961482048035, "step": 545 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.759765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2039.0, "completions/mean_length": 1708.005859375, "completions/mean_terminated_length": 632.7398071289062, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.18639583511137664, "frac_reward_zero_std": 0.65625, "grad_norm": 11.361701547499647, "learning_rate": 9.798720922042316e-07, "loss": 0.0209, "num_tokens": 518006945.0, "reward": 0.08203125, "reward_std": 0.13616865873336792, "rewards/accuracy_reward/mean": 0.08203125, "rewards/accuracy_reward/std": 0.2746807038784027, "step": 546 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.0, "completions/mean_length": 1875.9765625, "completions/mean_terminated_length": 752.7647094726562, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.18673721942476743, "frac_reward_zero_std": 0.875, "grad_norm": 3.1463425398936, "learning_rate": 9.7971324353692e-07, "loss": 0.0041, "num_tokens": 519044741.0, "reward": 0.01953125, "reward_std": 0.03944835811853409, "rewards/accuracy_reward/mean": 0.01953125, "rewards/accuracy_reward/std": 0.1385180652141571, "step": 547 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.76953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1905.0, "completions/mean_length": 1769.203125, "completions/mean_terminated_length": 838.3051147460938, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.18707860373815824, "frac_reward_zero_std": 0.78125, "grad_norm": 8.516965111440346, "learning_rate": 9.795537849702546e-07, "loss": 0.0167, "num_tokens": 520035997.0, "reward": 0.072265625, "reward_std": 0.08483455330133438, "rewards/accuracy_reward/mean": 0.072265625, "rewards/accuracy_reward/std": 0.2591804563999176, "step": 548 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.85546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1972.0, "completions/mean_length": 1850.822265625, "completions/mean_terminated_length": 683.7432861328125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.18741998805154902, "frac_reward_zero_std": 0.84375, "grad_norm": 23.723546969158985, "learning_rate": 9.79393716730558e-07, "loss": 0.008, "num_tokens": 521052338.0, "reward": 0.013671875, "reward_std": 0.043847277760505676, "rewards/accuracy_reward/mean": 0.01411290280520916, "rewards/accuracy_reward/std": 0.11807556450366974, "step": 549 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2031.0, "completions/mean_length": 1738.033203125, "completions/mean_terminated_length": 679.8706665039062, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.18776137236493984, "frac_reward_zero_std": 0.78125, "grad_norm": 5.554461766217026, "learning_rate": 9.792330390450179e-07, "loss": 0.0264, "num_tokens": 522016707.0, "reward": 0.05859375, "reward_std": 0.08659161627292633, "rewards/accuracy_reward/mean": 0.05859375, "rewards/accuracy_reward/std": 0.23509246110916138, "step": 550 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1986.0, "completions/mean_length": 1748.271484375, "completions/mean_terminated_length": 627.0648193359375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.18810275667833062, "frac_reward_zero_std": 0.875, "grad_norm": 19.822015894176953, "learning_rate": 9.790717521416865e-07, "loss": 0.0079, "num_tokens": 522992606.0, "reward": 0.013671875, "reward_std": 0.03741292655467987, "rewards/accuracy_reward/mean": 0.013671875, "rewards/accuracy_reward/std": 0.1162383034825325, "step": 551 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.818359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1983.0, "completions/mean_length": 1803.982421875, "completions/mean_terminated_length": 704.5913696289062, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.18844414099172144, "frac_reward_zero_std": 0.84375, "grad_norm": 7.063037350430578, "learning_rate": 9.789098562494813e-07, "loss": 0.0139, "num_tokens": 523985925.0, "reward": 0.037109375, "reward_std": 0.06271954625844955, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 552 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.861328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1869.0, "completions/mean_length": 1872.84765625, "completions/mean_terminated_length": 784.9295654296875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.18878552530511222, "frac_reward_zero_std": 0.84375, "grad_norm": 12.666618075475336, "learning_rate": 9.787473515981837e-07, "loss": 0.0037, "num_tokens": 525020919.0, "reward": 0.05078125, "reward_std": 0.06342554092407227, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 553 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.84765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1925.0, "completions/mean_length": 1857.02734375, "completions/mean_terminated_length": 794.4359130859375, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.18912690961850304, "frac_reward_zero_std": 0.90625, "grad_norm": 8.84247693319475, "learning_rate": 9.785842384184396e-07, "loss": 0.0031, "num_tokens": 526051141.0, "reward": 0.01171875, "reward_std": 0.032021719962358475, "rewards/accuracy_reward/mean": 0.01171875, "rewards/accuracy_reward/std": 0.10772226005792618, "step": 554 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.83203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1854.0, "completions/mean_length": 1805.044921875, "completions/mean_terminated_length": 602.1162719726562, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.18946829393189382, "frac_reward_zero_std": 0.84375, "grad_norm": 10.738478543181028, "learning_rate": 9.784205169417582e-07, "loss": 0.0077, "num_tokens": 527051596.0, "reward": 0.029296875, "reward_std": 0.05782270431518555, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 555 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.896484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1901.0, "completions/mean_length": 1913.193359375, "completions/mean_terminated_length": 745.7169799804688, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.18980967824528464, "frac_reward_zero_std": 0.96875, "grad_norm": 0.3147091493933924, "learning_rate": 9.782561874005121e-07, "loss": 0.0044, "num_tokens": 528110431.0, "reward": 0.00390625, "reward_std": 0.010673906654119492, "rewards/accuracy_reward/mean": 0.00390625, "rewards/accuracy_reward/std": 0.06243881583213806, "step": 556 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.869140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1948.0, "completions/mean_length": 1893.43359375, "completions/mean_terminated_length": 866.8358154296875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.19015106255867542, "frac_reward_zero_std": 0.75, "grad_norm": 28.941065225280187, "learning_rate": 9.78091250027937e-07, "loss": 0.0023, "num_tokens": 529154205.0, "reward": 0.05078125, "reward_std": 0.09145425260066986, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 557 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.845703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2020.0, "completions/mean_length": 1821.501953125, "completions/mean_terminated_length": 580.0632934570312, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.19049244687206623, "frac_reward_zero_std": 0.875, "grad_norm": 5.7571919857895075, "learning_rate": 9.779257050581316e-07, "loss": 0.0056, "num_tokens": 530163518.0, "reward": 0.013671875, "reward_std": 0.03741292655467987, "rewards/accuracy_reward/mean": 0.013671875, "rewards/accuracy_reward/std": 0.1162383034825325, "step": 558 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.91796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1542.0, "completions/mean_length": 1920.375, "completions/mean_terminated_length": 492.19049072265625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.19083383118545702, "frac_reward_zero_std": 1.0, "grad_norm": 0.0, "learning_rate": 9.777595527260567e-07, "loss": 0.0, "num_tokens": 531221182.0, "reward": 0.0, "reward_std": 0.0, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "step": 559 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.857421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2043.0, "completions/mean_length": 1834.982421875, "completions/mean_terminated_length": 553.9589233398438, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.19117521549884783, "frac_reward_zero_std": 0.90625, "grad_norm": 7.177725227789154, "learning_rate": 9.77592793267535e-07, "loss": 0.0041, "num_tokens": 532234117.0, "reward": 0.0234375, "reward_std": 0.03839729726314545, "rewards/accuracy_reward/mean": 0.0234375, "rewards/accuracy_reward/std": 0.15143637359142303, "step": 560 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.904296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1903.0, "completions/mean_length": 1890.23828125, "completions/mean_terminated_length": 399.551025390625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.19151659981223862, "frac_reward_zero_std": 0.96875, "grad_norm": 0.8671678512352751, "learning_rate": 9.774254269192506e-07, "loss": 0.0038, "num_tokens": 533274975.0, "reward": 0.001953125, "reward_std": 0.0078125, "rewards/accuracy_reward/mean": 0.001953125, "rewards/accuracy_reward/std": 0.04419417306780815, "step": 561 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.83203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1934.0, "completions/mean_length": 1822.31640625, "completions/mean_terminated_length": 704.3953247070312, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.19185798412562943, "frac_reward_zero_std": 0.875, "grad_norm": 7.537147527139063, "learning_rate": 9.772574539187503e-07, "loss": 0.011, "num_tokens": 534289729.0, "reward": 0.02734375, "reward_std": 0.04192390665411949, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 562 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.826171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2029.0, "completions/mean_length": 1845.525390625, "completions/mean_terminated_length": 883.2022705078125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.19219936843902022, "frac_reward_zero_std": 0.90625, "grad_norm": 4.534077515268373, "learning_rate": 9.770888745044405e-07, "loss": 0.0118, "num_tokens": 535310142.0, "reward": 0.021484375, "reward_std": 0.03773209825158119, "rewards/accuracy_reward/mean": 0.02217741869390011, "rewards/accuracy_reward/std": 0.14740893244743347, "step": 563 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.81640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2043.0, "completions/mean_length": 1808.1484375, "completions/mean_terminated_length": 741.574462890625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.19254075275241103, "frac_reward_zero_std": 0.84375, "grad_norm": 13.679005661288093, "learning_rate": 9.769196889155888e-07, "loss": 0.0043, "num_tokens": 536323082.0, "reward": 0.04296875, "reward_std": 0.04907120764255524, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 564 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.90625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1792.0, "completions/mean_length": 1914.419921875, "completions/mean_terminated_length": 623.1458740234375, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.19288213706580182, "frac_reward_zero_std": 0.90625, "grad_norm": 6.152320412280184, "learning_rate": 9.767498973923236e-07, "loss": 0.0057, "num_tokens": 537377937.0, "reward": 0.009765625, "reward_std": 0.029160313308238983, "rewards/accuracy_reward/mean": 0.009765625, "rewards/accuracy_reward/std": 0.09843364357948303, "step": 565 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2039.0, "completions/mean_length": 1796.42578125, "completions/mean_terminated_length": 759.9400024414062, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.19322352137919263, "frac_reward_zero_std": 0.875, "grad_norm": 12.836133321230195, "learning_rate": 9.765795001756326e-07, "loss": 0.0119, "num_tokens": 538380043.0, "reward": 0.044921875, "reward_std": 0.05133409798145294, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 566 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.880859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1944.0, "completions/mean_length": 1894.626953125, "completions/mean_terminated_length": 761.5573120117188, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.19356490569258342, "frac_reward_zero_std": 0.90625, "grad_norm": 9.405455281696756, "learning_rate": 9.764084975073635e-07, "loss": 0.005, "num_tokens": 539434492.0, "reward": 0.01171875, "reward_std": 0.02960042469203472, "rewards/accuracy_reward/mean": 0.01171875, "rewards/accuracy_reward/std": 0.10772226005792618, "step": 567 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.888671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1885.0, "completions/mean_length": 1888.25, "completions/mean_terminated_length": 613.0526123046875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.19390629000597423, "frac_reward_zero_std": 0.8125, "grad_norm": 8.965843961012798, "learning_rate": 9.762368896302234e-07, "loss": 0.0105, "num_tokens": 540474572.0, "reward": 0.01953125, "reward_std": 0.058320626616477966, "rewards/accuracy_reward/mean": 0.01953125, "rewards/accuracy_reward/std": 0.1385180652141571, "step": 568 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.87109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2004.0, "completions/mean_length": 1873.9140625, "completions/mean_terminated_length": 697.5151977539062, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.19424767431936502, "frac_reward_zero_std": 0.90625, "grad_norm": 4.310290168024493, "learning_rate": 9.760646767877784e-07, "loss": -0.0049, "num_tokens": 541506928.0, "reward": 0.01171875, "reward_std": 0.031083684414625168, "rewards/accuracy_reward/mean": 0.01171875, "rewards/accuracy_reward/std": 0.10772226005792618, "step": 569 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.884765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1956.0, "completions/mean_length": 1910.544921875, "completions/mean_terminated_length": 855.1694946289062, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.19458905863275583, "frac_reward_zero_std": 0.90625, "grad_norm": 4.06847169525495, "learning_rate": 9.758918592244528e-07, "loss": 0.0128, "num_tokens": 542558423.0, "reward": 0.017578125, "reward_std": 0.03536957502365112, "rewards/accuracy_reward/mean": 0.017578125, "rewards/accuracy_reward/std": 0.13154059648513794, "step": 570 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.806640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2037.0, "completions/mean_length": 1820.466796875, "completions/mean_terminated_length": 871.2626342773438, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.19493044294614661, "frac_reward_zero_std": 0.8125, "grad_norm": 12.81160156550978, "learning_rate": 9.757184371855298e-07, "loss": 0.0073, "num_tokens": 543575062.0, "reward": 0.04296875, "reward_std": 0.07415273785591125, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 571 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.841796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2005.0, "completions/mean_length": 1841.806640625, "completions/mean_terminated_length": 744.654296875, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.19527182725953743, "frac_reward_zero_std": 0.84375, "grad_norm": 15.483274572528053, "learning_rate": 9.7554441091715e-07, "loss": 0.0142, "num_tokens": 544594691.0, "reward": 0.04296875, "reward_std": 0.06037056818604469, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 572 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1845.0, "completions/mean_length": 1849.029296875, "completions/mean_terminated_length": 835.2261962890625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.19561321157292821, "frac_reward_zero_std": 0.96875, "grad_norm": 1.1304099438143267, "learning_rate": 9.753697806663124e-07, "loss": 0.0001, "num_tokens": 545617922.0, "reward": 0.00390625, "reward_std": 0.010673906654119492, "rewards/accuracy_reward/mean": 0.00390625, "rewards/accuracy_reward/std": 0.06243881583213806, "step": 573 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.810546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1913.0, "completions/mean_length": 1766.611328125, "completions/mean_terminated_length": 562.73193359375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.19595459588631903, "frac_reward_zero_std": 0.875, "grad_norm": 20.944897509612453, "learning_rate": 9.75194546680872e-07, "loss": -0.0149, "num_tokens": 546599259.0, "reward": 0.03125, "reward_std": 0.05012226849794388, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 574 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.91796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1998.0, "completions/mean_length": 1935.37890625, "completions/mean_terminated_length": 675.0952758789062, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.1962959801997098, "frac_reward_zero_std": 0.90625, "grad_norm": 2.175657373291508, "learning_rate": 9.750187092095422e-07, "loss": 0.0063, "num_tokens": 547664253.0, "reward": 0.021484375, "reward_std": 0.036420635879039764, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 575 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1939.0, "completions/mean_length": 1804.53125, "completions/mean_terminated_length": 631.45458984375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.19663736451310063, "frac_reward_zero_std": 0.9375, "grad_norm": 4.947413235457389, "learning_rate": 9.748422685018911e-07, "loss": 0.0011, "num_tokens": 548673133.0, "reward": 0.00390625, "reward_std": 0.015625, "rewards/accuracy_reward/mean": 0.004032257944345474, "rewards/accuracy_reward/std": 0.06343588978052139, "step": 576 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.853515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2039.0, "completions/mean_length": 1867.91015625, "completions/mean_terminated_length": 818.586669921875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.1969787488264914, "frac_reward_zero_std": 0.9375, "grad_norm": 4.8444711746001765, "learning_rate": 9.74665224808345e-07, "loss": -0.0008, "num_tokens": 549709295.0, "reward": 0.015625, "reward_std": 0.023823359981179237, "rewards/accuracy_reward/mean": 0.015625, "rewards/accuracy_reward/std": 0.12414088100194931, "step": 577 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.759765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2029.0, "completions/mean_length": 1764.1484375, "completions/mean_terminated_length": 866.43896484375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.19732013313988223, "frac_reward_zero_std": 0.84375, "grad_norm": 11.501801099434747, "learning_rate": 9.744875783801844e-07, "loss": -0.0039, "num_tokens": 550688971.0, "reward": 0.0390625, "reward_std": 0.06299334019422531, "rewards/accuracy_reward/mean": 0.04032257944345474, "rewards/accuracy_reward/std": 0.19691328704357147, "step": 578 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1974.0, "completions/mean_length": 1819.109375, "completions/mean_terminated_length": 827.25, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.197661517453273, "frac_reward_zero_std": 0.90625, "grad_norm": 4.010604069433794, "learning_rate": 9.743093294695461e-07, "loss": 0.0123, "num_tokens": 551699731.0, "reward": 0.01953125, "reward_std": 0.036034777760505676, "rewards/accuracy_reward/mean": 0.01953125, "rewards/accuracy_reward/std": 0.1385180652141571, "step": 579 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.791015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1951.0, "completions/mean_length": 1810.64453125, "completions/mean_terminated_length": 912.2429809570312, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.19800290176666382, "frac_reward_zero_std": 0.8125, "grad_norm": 7.29025529971875, "learning_rate": 9.741304783294218e-07, "loss": 0.022, "num_tokens": 552705757.0, "reward": 0.041015625, "reward_std": 0.07339344918727875, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 580 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.806640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2015.0, "completions/mean_length": 1795.353515625, "completions/mean_terminated_length": 741.8485107421875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.1983442860800546, "frac_reward_zero_std": 0.90625, "grad_norm": 10.37541462003429, "learning_rate": 9.739510252136584e-07, "loss": 0.0041, "num_tokens": 553706850.0, "reward": 0.033203125, "reward_std": 0.029160313308238983, "rewards/accuracy_reward/mean": 0.033203125, "rewards/accuracy_reward/std": 0.17934183776378632, "step": 581 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1966.0, "completions/mean_length": 1823.11328125, "completions/mean_terminated_length": 796.45654296875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.19868567039344542, "frac_reward_zero_std": 0.84375, "grad_norm": 15.049323590650918, "learning_rate": 9.737709703769562e-07, "loss": 0.0186, "num_tokens": 554726348.0, "reward": 0.03515625, "reward_std": 0.05507335811853409, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 582 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.91796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1946.0, "completions/mean_length": 1936.390625, "completions/mean_terminated_length": 687.4285888671875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.1990270547068362, "frac_reward_zero_std": 0.875, "grad_norm": 7.8096486926393, "learning_rate": 9.735903140748702e-07, "loss": 0.0106, "num_tokens": 555789172.0, "reward": 0.01171875, "reward_std": 0.036034777760505676, "rewards/accuracy_reward/mean": 0.01171875, "rewards/accuracy_reward/std": 0.10772226005792618, "step": 583 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.89453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1929.0, "completions/mean_length": 1915.69921875, "completions/mean_terminated_length": 793.5925903320312, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.19936843902022702, "frac_reward_zero_std": 0.96875, "grad_norm": 3.2684767544500843, "learning_rate": 9.734090565638092e-07, "loss": -0.0003, "num_tokens": 556851962.0, "reward": 0.001953125, "reward_std": 0.0078125, "rewards/accuracy_reward/mean": 0.001953125, "rewards/accuracy_reward/std": 0.04419417306780815, "step": 584 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.857421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2028.0, "completions/mean_length": 1867.755859375, "completions/mean_terminated_length": 783.8218994140625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.1997098233336178, "frac_reward_zero_std": 0.84375, "grad_norm": 5.368717374353762, "learning_rate": 9.73227198101035e-07, "loss": 0.0052, "num_tokens": 557877565.0, "reward": 0.029296875, "reward_std": 0.05298367142677307, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 585 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1894.0, "completions/mean_length": 1888.966796875, "completions/mean_terminated_length": 775.734375, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.20005120764700862, "frac_reward_zero_std": 0.96875, "grad_norm": 0.9319445263626492, "learning_rate": 9.730447389446623e-07, "loss": 0.0007, "num_tokens": 558922700.0, "reward": 0.001953125, "reward_std": 0.0078125, "rewards/accuracy_reward/mean": 0.001953125, "rewards/accuracy_reward/std": 0.04419417306780815, "step": 586 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.87109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1855.0, "completions/mean_length": 1903.37109375, "completions/mean_terminated_length": 926.8485107421875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.2003925919603994, "frac_reward_zero_std": 0.9375, "grad_norm": 3.7739725416482286, "learning_rate": 9.728616793536587e-07, "loss": 0.0051, "num_tokens": 559979466.0, "reward": 0.0078125, "reward_std": 0.020409777760505676, "rewards/accuracy_reward/mean": 0.0078125, "rewards/accuracy_reward/std": 0.08812850713729858, "step": 587 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.90625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1786.0, "completions/mean_length": 1913.91796875, "completions/mean_terminated_length": 617.7916870117188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.20073397627379022, "frac_reward_zero_std": 0.9375, "grad_norm": 0.4087371599173673, "learning_rate": 9.726780195878438e-07, "loss": 0.0125, "num_tokens": 561035120.0, "reward": 0.009765625, "reward_std": 0.02178792469203472, "rewards/accuracy_reward/mean": 0.009765625, "rewards/accuracy_reward/std": 0.09843364357948303, "step": 588 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1880.0, "completions/mean_length": 1915.935546875, "completions/mean_terminated_length": 747.673095703125, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.201075360587181, "frac_reward_zero_std": 0.875, "grad_norm": 26.011481222330602, "learning_rate": 9.724937599078888e-07, "loss": 0.0022, "num_tokens": 562094127.0, "reward": 0.01171875, "reward_std": 0.036034777760505676, "rewards/accuracy_reward/mean": 0.012096773833036423, "rewards/accuracy_reward/std": 0.10942844301462173, "step": 589 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1539.0, "completions/mean_length": 1928.267578125, "completions/mean_terminated_length": 515.4249877929688, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.20141674490057182, "frac_reward_zero_std": 0.9375, "grad_norm": 3.126644533668479, "learning_rate": 9.72308900575317e-07, "loss": 0.0209, "num_tokens": 563148872.0, "reward": 0.015625, "reward_std": 0.027950849384069443, "rewards/accuracy_reward/mean": 0.015625, "rewards/accuracy_reward/std": 0.12414088100194931, "step": 590 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.84375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2041.0, "completions/mean_length": 1844.841796875, "completions/mean_terminated_length": 747.7875366210938, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.2017581292139626, "frac_reward_zero_std": 0.84375, "grad_norm": 14.645977451419562, "learning_rate": 9.72123441852502e-07, "loss": 0.0128, "num_tokens": 564175095.0, "reward": 0.02734375, "reward_std": 0.05311024188995361, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 591 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.904296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1967.0, "completions/mean_length": 1915.9296875, "completions/mean_terminated_length": 668.0, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.20209951352735342, "frac_reward_zero_std": 0.90625, "grad_norm": 11.592869084790411, "learning_rate": 9.719373840026686e-07, "loss": 0.0107, "num_tokens": 565229731.0, "reward": 0.03125, "reward_std": 0.04369450733065605, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 592 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.814453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2031.0, "completions/mean_length": 1782.19140625, "completions/mean_terminated_length": 615.4315795898438, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.2024408978407442, "frac_reward_zero_std": 0.84375, "grad_norm": 21.605193368541222, "learning_rate": 9.717507272898922e-07, "loss": 0.0149, "num_tokens": 566223957.0, "reward": 0.029296875, "reward_std": 0.06150216981768608, "rewards/accuracy_reward/mean": 0.030241934582591057, "rewards/accuracy_reward/std": 0.1714252382516861, "step": 593 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1968.0, "completions/mean_length": 1821.900390625, "completions/mean_terminated_length": 440.1805725097656, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.20278228215413502, "frac_reward_zero_std": 0.78125, "grad_norm": 6.0613364526902025, "learning_rate": 9.715634719790978e-07, "loss": 0.022, "num_tokens": 567227362.0, "reward": 0.046875, "reward_std": 0.08670367300510406, "rewards/accuracy_reward/mean": 0.046875, "rewards/accuracy_reward/std": 0.21157780289649963, "step": 594 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.91015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2021.0, "completions/mean_length": 1946.060546875, "completions/mean_terminated_length": 913.3695678710938, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.2031236664675258, "frac_reward_zero_std": 0.90625, "grad_norm": 4.70812589186284, "learning_rate": 9.713756183360597e-07, "loss": 0.0033, "num_tokens": 568298561.0, "reward": 0.013671875, "reward_std": 0.030584799125790596, "rewards/accuracy_reward/mean": 0.013671875, "rewards/accuracy_reward/std": 0.1162383034825325, "step": 595 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.861328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2000.0, "completions/mean_length": 1868.12890625, "completions/mean_terminated_length": 750.9013671875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.20346505078091662, "frac_reward_zero_std": 0.875, "grad_norm": 14.122825166578655, "learning_rate": 9.711871666274021e-07, "loss": 0.0003, "num_tokens": 569320995.0, "reward": 0.03515625, "reward_std": 0.04505910724401474, "rewards/accuracy_reward/mean": 0.03629032149910927, "rewards/accuracy_reward/std": 0.1872003972530365, "step": 596 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.87890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1893.0, "completions/mean_length": 1896.140625, "completions/mean_terminated_length": 793.9354858398438, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.2038064350943074, "frac_reward_zero_std": 0.9375, "grad_norm": 11.582825324714353, "learning_rate": 9.709981171205977e-07, "loss": -0.0005, "num_tokens": 570363723.0, "reward": 0.01171875, "reward_std": 0.022772299125790596, "rewards/accuracy_reward/mean": 0.012096773833036423, "rewards/accuracy_reward/std": 0.10942844301462173, "step": 597 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1984.0, "completions/mean_length": 1825.79296875, "completions/mean_terminated_length": 755.1591186523438, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.20414781940769822, "frac_reward_zero_std": 0.78125, "grad_norm": 6.968227355568104, "learning_rate": 9.708084700839678e-07, "loss": 0.0283, "num_tokens": 571380961.0, "reward": 0.076171875, "reward_std": 0.09793634712696075, "rewards/accuracy_reward/mean": 0.076171875, "rewards/accuracy_reward/std": 0.26553234457969666, "step": 598 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.84765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2022.0, "completions/mean_length": 1874.10546875, "completions/mean_terminated_length": 907.1154174804688, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.204489203721089, "frac_reward_zero_std": 0.875, "grad_norm": 8.123197532296599, "learning_rate": 9.706182257866812e-07, "loss": 0.011, "num_tokens": 572426135.0, "reward": 0.0390625, "reward_std": 0.046542368829250336, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 599 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.798828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1915.0, "completions/mean_length": 1784.28515625, "completions/mean_terminated_length": 737.1068115234375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.20483058803447982, "frac_reward_zero_std": 0.78125, "grad_norm": 24.643624675033465, "learning_rate": 9.704273844987555e-07, "loss": 0.0078, "num_tokens": 573419657.0, "reward": 0.064453125, "reward_std": 0.08780106902122498, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 600 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.78515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2034.0, "completions/mean_length": 1817.62109375, "completions/mean_terminated_length": 975.6908569335938, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.20517197234787063, "frac_reward_zero_std": 0.71875, "grad_norm": 19.13542093834099, "learning_rate": 9.702359464910546e-07, "loss": 0.0197, "num_tokens": 574426407.0, "reward": 0.072265625, "reward_std": 0.11113909631967545, "rewards/accuracy_reward/mean": 0.072265625, "rewards/accuracy_reward/std": 0.2591804563999176, "step": 601 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.755859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1928.0, "completions/mean_length": 1709.392578125, "completions/mean_terminated_length": 661.0640258789062, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.20551335666126141, "frac_reward_zero_std": 0.875, "grad_norm": 7.179275537858811, "learning_rate": 9.700439120352898e-07, "loss": 0.0017, "num_tokens": 575380544.0, "reward": 0.021484375, "reward_std": 0.04604348540306091, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 602 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.85546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1944.0, "completions/mean_length": 1863.49609375, "completions/mean_terminated_length": 771.4324340820312, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.20585474097465223, "frac_reward_zero_std": 0.78125, "grad_norm": 6.2513676266546545, "learning_rate": 9.69851281404019e-07, "loss": 0.0227, "num_tokens": 576409502.0, "reward": 0.02734375, "reward_std": 0.07135801017284393, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 603 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.78125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2021.0, "completions/mean_length": 1762.978515625, "completions/mean_terminated_length": 745.044677734375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.20619612528804301, "frac_reward_zero_std": 0.8125, "grad_norm": 4.460169727717185, "learning_rate": 9.696580548706462e-07, "loss": 0.0157, "num_tokens": 577391123.0, "reward": 0.029296875, "reward_std": 0.06092274188995361, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 604 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.76953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2024.0, "completions/mean_length": 1777.99609375, "completions/mean_terminated_length": 876.4576416015625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.20653750960143383, "frac_reward_zero_std": 0.75, "grad_norm": 2.8758813445490032, "learning_rate": 9.69464232709421e-07, "loss": 0.0214, "num_tokens": 578374577.0, "reward": 0.046875, "reward_std": 0.0876617580652237, "rewards/accuracy_reward/mean": 0.046875, "rewards/accuracy_reward/std": 0.21157780289649963, "step": 605 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.814453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1995.0, "completions/mean_length": 1781.177734375, "completions/mean_terminated_length": 609.9684448242188, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.2068788939148246, "frac_reward_zero_std": 0.96875, "grad_norm": 1.0850371593350496, "learning_rate": 9.692698151954383e-07, "loss": -0.0002, "num_tokens": 579364316.0, "reward": 0.0078125, "reward_std": 0.013975424692034721, "rewards/accuracy_reward/mean": 0.0078125, "rewards/accuracy_reward/std": 0.08812850713729858, "step": 606 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.830078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2040.0, "completions/mean_length": 1839.98828125, "completions/mean_terminated_length": 823.8390502929688, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.20722027822821543, "frac_reward_zero_std": 0.9375, "grad_norm": 10.426151879229034, "learning_rate": 9.690748026046386e-07, "loss": -0.0004, "num_tokens": 580380614.0, "reward": 0.005859375, "reward_std": 0.01848640665411949, "rewards/accuracy_reward/mean": 0.005859375, "rewards/accuracy_reward/std": 0.07639661431312561, "step": 607 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.76171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2017.0, "completions/mean_length": 1716.9140625, "completions/mean_terminated_length": 658.5245361328125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.2075616625416062, "frac_reward_zero_std": 0.875, "grad_norm": 2.8669272294840247, "learning_rate": 9.688791952138068e-07, "loss": -0.0039, "num_tokens": 581339354.0, "reward": 0.048828125, "reward_std": 0.04230976849794388, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 608 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.740234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2038.0, "completions/mean_length": 1751.7421875, "completions/mean_terminated_length": 907.518798828125, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.20790304685499703, "frac_reward_zero_std": 0.9375, "grad_norm": 3.1731800354683215, "learning_rate": 9.686829933005709e-07, "loss": 0.0035, "num_tokens": 582312998.0, "reward": 0.005859375, "reward_std": 0.01848640665411949, "rewards/accuracy_reward/mean": 0.005859375, "rewards/accuracy_reward/std": 0.07639661431312561, "step": 609 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2000.0, "completions/mean_length": 1674.75, "completions/mean_terminated_length": 823.2628173828125, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.2082444311683878, "frac_reward_zero_std": 0.78125, "grad_norm": 9.574243232286076, "learning_rate": 9.684861971434043e-07, "loss": 0.0001, "num_tokens": 583257558.0, "reward": 0.05078125, "reward_std": 0.08263043314218521, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 610 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.775390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1938.0, "completions/mean_length": 1743.79296875, "completions/mean_terminated_length": 693.6173706054688, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.20858581548177862, "frac_reward_zero_std": 0.8125, "grad_norm": 22.244181455949896, "learning_rate": 9.682888070216231e-07, "loss": 0.0003, "num_tokens": 584229500.0, "reward": 0.037109375, "reward_std": 0.06695909798145294, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 611 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1917.0, "completions/mean_length": 1804.123046875, "completions/mean_terminated_length": 847.8750610351562, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.2089271997951694, "frac_reward_zero_std": 0.90625, "grad_norm": 9.826526998632836, "learning_rate": 9.680908232153865e-07, "loss": 0.0027, "num_tokens": 585237403.0, "reward": 0.009765625, "reward_std": 0.028222277760505676, "rewards/accuracy_reward/mean": 0.009765625, "rewards/accuracy_reward/std": 0.09843364357948303, "step": 612 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1887.0, "completions/mean_length": 1759.572265625, "completions/mean_terminated_length": 774.9396362304688, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.20926858410856022, "frac_reward_zero_std": 0.875, "grad_norm": 2.4591367622231552, "learning_rate": 9.67892246005696e-07, "loss": -0.0071, "num_tokens": 586210032.0, "reward": 0.029296875, "reward_std": 0.04859926179051399, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 613 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.76171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1988.0, "completions/mean_length": 1724.783203125, "completions/mean_terminated_length": 691.5491333007812, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.209609968421951, "frac_reward_zero_std": 0.8125, "grad_norm": 15.078602201675519, "learning_rate": 9.67693075674396e-07, "loss": 0.028, "num_tokens": 587171249.0, "reward": 0.080078125, "reward_std": 0.08311038464307785, "rewards/accuracy_reward/mean": 0.080078125, "rewards/accuracy_reward/std": 0.271679550409317, "step": 614 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2032.0, "completions/mean_length": 1770.412109375, "completions/mean_terminated_length": 681.9615478515625, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.20995135273534182, "frac_reward_zero_std": 0.84375, "grad_norm": 7.930803448566627, "learning_rate": 9.674933125041722e-07, "loss": -0.0017, "num_tokens": 588159108.0, "reward": 0.0234375, "reward_std": 0.054249756038188934, "rewards/accuracy_reward/mean": 0.0234375, "rewards/accuracy_reward/std": 0.15143637359142303, "step": 615 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.669921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2038.0, "completions/mean_length": 1642.30078125, "completions/mean_terminated_length": 818.8994140625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.2102927370487326, "frac_reward_zero_std": 0.90625, "grad_norm": 5.684390193283198, "learning_rate": 9.672929567785517e-07, "loss": 0.0016, "num_tokens": 589083790.0, "reward": 0.0390625, "reward_std": 0.04318207502365112, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 616 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.70703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2037.0, "completions/mean_length": 1675.79296875, "completions/mean_terminated_length": 777.5333251953125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.21063412136212342, "frac_reward_zero_std": 0.65625, "grad_norm": 14.299613801039989, "learning_rate": 9.67092008781903e-07, "loss": 0.0074, "num_tokens": 590024660.0, "reward": 0.04296875, "reward_std": 0.10178204625844955, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 617 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.794921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1988.0, "completions/mean_length": 1775.001953125, "completions/mean_terminated_length": 716.8095703125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.2109755056755142, "frac_reward_zero_std": 0.8125, "grad_norm": 4.809029334339867, "learning_rate": 9.668904687994351e-07, "loss": 0.0202, "num_tokens": 591002597.0, "reward": 0.037109375, "reward_std": 0.06587383151054382, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 618 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.697265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2007.0, "completions/mean_length": 1685.38671875, "completions/mean_terminated_length": 850.2064208984375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.21131688998890502, "frac_reward_zero_std": 0.9375, "grad_norm": 5.840188475369804, "learning_rate": 9.66688337117197e-07, "loss": -0.0032, "num_tokens": 591943883.0, "reward": 0.029296875, "reward_std": 0.030584799125790596, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 619 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2026.0, "completions/mean_length": 1631.392578125, "completions/mean_terminated_length": 714.8562622070312, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.2116582743022958, "frac_reward_zero_std": 0.875, "grad_norm": 5.278747550405541, "learning_rate": 9.664856140220778e-07, "loss": -0.0132, "num_tokens": 592856628.0, "reward": 0.033203125, "reward_std": 0.04219770431518555, "rewards/accuracy_reward/mean": 0.033203125, "rewards/accuracy_reward/std": 0.17934183776378632, "step": 620 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.720703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2037.0, "completions/mean_length": 1716.3125, "completions/mean_terminated_length": 860.4195556640625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.21199965861568662, "frac_reward_zero_std": 0.90625, "grad_norm": 11.36693791810471, "learning_rate": 9.662822998018056e-07, "loss": -0.0016, "num_tokens": 593813252.0, "reward": 0.029296875, "reward_std": 0.04456022381782532, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 621 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.794921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1932.0, "completions/mean_length": 1812.8203125, "completions/mean_terminated_length": 901.2190551757812, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.2123410429290774, "frac_reward_zero_std": 0.71875, "grad_norm": 10.517391880169027, "learning_rate": 9.66078394744948e-07, "loss": 0.0232, "num_tokens": 594819560.0, "reward": 0.0546875, "reward_std": 0.10519562661647797, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 622 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.802734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1925.0, "completions/mean_length": 1793.93359375, "completions/mean_terminated_length": 760.0593872070312, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.21268242724246822, "frac_reward_zero_std": 0.78125, "grad_norm": 5.878379186252776, "learning_rate": 9.65873899140911e-07, "loss": 0.017, "num_tokens": 595811814.0, "reward": 0.046875, "reward_std": 0.08675792813301086, "rewards/accuracy_reward/mean": 0.046875, "rewards/accuracy_reward/std": 0.21157780289649963, "step": 623 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.75390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2022.0, "completions/mean_length": 1720.763671875, "completions/mean_terminated_length": 718.6111450195312, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.213023811555859, "frac_reward_zero_std": 0.84375, "grad_norm": 6.563406905558506, "learning_rate": 9.656688132799382e-07, "loss": 0.0064, "num_tokens": 596782269.0, "reward": 0.037109375, "reward_std": 0.05335709825158119, "rewards/accuracy_reward/mean": 0.038306452333927155, "rewards/accuracy_reward/std": 0.19212885200977325, "step": 624 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.85546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2039.0, "completions/mean_length": 1875.384765625, "completions/mean_terminated_length": 854.4324340820312, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.21336519586924982, "frac_reward_zero_std": 0.90625, "grad_norm": 7.929407269197637, "learning_rate": 9.65463137453112e-07, "loss": -0.0004, "num_tokens": 597821186.0, "reward": 0.0234375, "reward_std": 0.03779878467321396, "rewards/accuracy_reward/mean": 0.0234375, "rewards/accuracy_reward/std": 0.15143637359142303, "step": 625 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2015.0, "completions/mean_length": 1852.630859375, "completions/mean_terminated_length": 857.1785888671875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.2137065801826406, "frac_reward_zero_std": 0.78125, "grad_norm": 3.607552224731443, "learning_rate": 9.652568719523516e-07, "loss": 0.0235, "num_tokens": 598849925.0, "reward": 0.0390625, "reward_std": 0.08109389245510101, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 626 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1947.0, "completions/mean_length": 1803.330078125, "completions/mean_terminated_length": 743.09375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.21404796449603142, "frac_reward_zero_std": 0.84375, "grad_norm": 5.434637836004345, "learning_rate": 9.650500170704127e-07, "loss": 0.0199, "num_tokens": 599851790.0, "reward": 0.037109375, "reward_std": 0.06118202954530716, "rewards/accuracy_reward/mean": 0.038306452333927155, "rewards/accuracy_reward/std": 0.19212883710861206, "step": 627 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.791015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2009.0, "completions/mean_length": 1785.666015625, "completions/mean_terminated_length": 792.7196044921875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.2143893488094222, "frac_reward_zero_std": 0.84375, "grad_norm": 7.556227191545887, "learning_rate": 9.648425731008884e-07, "loss": 0.014, "num_tokens": 600845667.0, "reward": 0.0390625, "reward_std": 0.06449805945158005, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 628 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.869140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1654.0, "completions/mean_length": 1856.419921875, "completions/mean_terminated_length": 583.9850463867188, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.21473073312281302, "frac_reward_zero_std": 0.90625, "grad_norm": 1.6683977493428692, "learning_rate": 9.646345403382073e-07, "loss": 0.0144, "num_tokens": 601879242.0, "reward": 0.01171875, "reward_std": 0.032021719962358475, "rewards/accuracy_reward/mean": 0.01171875, "rewards/accuracy_reward/std": 0.10772226005792618, "step": 629 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.865234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1831.0, "completions/mean_length": 1883.33203125, "completions/mean_terminated_length": 826.115966796875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.2150721174362038, "frac_reward_zero_std": 0.9375, "grad_norm": 10.02712140385809, "learning_rate": 9.644259190776339e-07, "loss": 0.0041, "num_tokens": 602916468.0, "reward": 0.01953125, "reward_std": 0.028608137741684914, "rewards/accuracy_reward/mean": 0.01953125, "rewards/accuracy_reward/std": 0.1385180652141571, "step": 630 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2003.0, "completions/mean_length": 1929.96875, "completions/mean_terminated_length": 885.84619140625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.21541350174959462, "frac_reward_zero_std": 0.875, "grad_norm": 2.838934611715861, "learning_rate": 9.642167096152678e-07, "loss": 0.0077, "num_tokens": 603979028.0, "reward": 0.009765625, "reward_std": 0.03411140665411949, "rewards/accuracy_reward/mean": 0.009765625, "rewards/accuracy_reward/std": 0.09843364357948303, "step": 631 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.76953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1989.0, "completions/mean_length": 1762.919921875, "completions/mean_terminated_length": 811.0423583984375, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.2157548860629854, "frac_reward_zero_std": 0.625, "grad_norm": 14.619106996182678, "learning_rate": 9.640069122480437e-07, "loss": 0.0278, "num_tokens": 604958843.0, "reward": 0.07421875, "reward_std": 0.12224511802196503, "rewards/accuracy_reward/mean": 0.07421875, "rewards/accuracy_reward/std": 0.2623828947544098, "step": 632 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.779296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2010.0, "completions/mean_length": 1786.275390625, "completions/mean_terminated_length": 862.1327514648438, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.21609627037637621, "frac_reward_zero_std": 0.90625, "grad_norm": 9.271328246169466, "learning_rate": 9.637965272737305e-07, "loss": 0.0076, "num_tokens": 605946920.0, "reward": 0.017578125, "reward_std": 0.03724660724401474, "rewards/accuracy_reward/mean": 0.017578125, "rewards/accuracy_reward/std": 0.13154059648513794, "step": 633 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.783203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2041.0, "completions/mean_length": 1765.341796875, "completions/mean_terminated_length": 744.2072143554688, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.216437654689767, "frac_reward_zero_std": 0.84375, "grad_norm": 5.266560544683273, "learning_rate": 9.635855549909314e-07, "loss": 0.0048, "num_tokens": 606923639.0, "reward": 0.03515625, "reward_std": 0.0645298957824707, "rewards/accuracy_reward/mean": 0.03629032149910927, "rewards/accuracy_reward/std": 0.1872003823518753, "step": 634 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.869140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2044.0, "completions/mean_length": 1895.080078125, "completions/mean_terminated_length": 879.4179077148438, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.21677903900315781, "frac_reward_zero_std": 0.9375, "grad_norm": 1.9201946016079303, "learning_rate": 9.63373995699083e-07, "loss": -0.0006, "num_tokens": 607973200.0, "reward": 0.0078125, "reward_std": 0.020409777760505676, "rewards/accuracy_reward/mean": 0.0078125, "rewards/accuracy_reward/std": 0.08812850713729858, "step": 635 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.876953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2029.0, "completions/mean_length": 1909.587890625, "completions/mean_terminated_length": 923.1270141601562, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.2171204233165486, "frac_reward_zero_std": 0.96875, "grad_norm": 0.06175537020243458, "learning_rate": 9.631618496984546e-07, "loss": 0.0054, "num_tokens": 609026973.0, "reward": 0.001953125, "reward_std": 0.0078125, "rewards/accuracy_reward/mean": 0.001953125, "rewards/accuracy_reward/std": 0.04419417306780815, "step": 636 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.837890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1997.0, "completions/mean_length": 1863.95703125, "completions/mean_terminated_length": 912.6987915039062, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.2174618076299394, "frac_reward_zero_std": 0.9375, "grad_norm": 3.7033716440696702, "learning_rate": 9.62949117290149e-07, "loss": 0.0071, "num_tokens": 610061111.0, "reward": 0.015625, "reward_std": 0.02629890665411949, "rewards/accuracy_reward/mean": 0.015625, "rewards/accuracy_reward/std": 0.12414088100194931, "step": 637 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1986.0, "completions/mean_length": 1822.32421875, "completions/mean_terminated_length": 936.9807739257812, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.2178031919433302, "frac_reward_zero_std": 0.84375, "grad_norm": 7.692842122644058, "learning_rate": 9.627357987761007e-07, "loss": 0.0192, "num_tokens": 611062637.0, "reward": 0.0625, "reward_std": 0.062220655381679535, "rewards/accuracy_reward/mean": 0.06451612710952759, "rewards/accuracy_reward/std": 0.2459181249141693, "step": 638 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1896.0, "completions/mean_length": 1789.890625, "completions/mean_terminated_length": 777.3077392578125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.218144576256721, "frac_reward_zero_std": 0.78125, "grad_norm": 4.649960745861791, "learning_rate": 9.625218944590763e-07, "loss": 0.0114, "num_tokens": 612052261.0, "reward": 0.029296875, "reward_std": 0.07108421623706818, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 639 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.775390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2017.0, "completions/mean_length": 1768.314453125, "completions/mean_terminated_length": 802.791259765625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.2184859605701118, "frac_reward_zero_std": 0.84375, "grad_norm": 10.403190356293088, "learning_rate": 9.623074046426744e-07, "loss": -0.0029, "num_tokens": 613033414.0, "reward": 0.033203125, "reward_std": 0.048086829483509064, "rewards/accuracy_reward/mean": 0.033203125, "rewards/accuracy_reward/std": 0.17934183776378632, "step": 640 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.724609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2038.0, "completions/mean_length": 1727.623046875, "completions/mean_terminated_length": 884.6453857421875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.2188273448835026, "frac_reward_zero_std": 0.84375, "grad_norm": 3.370851484601525, "learning_rate": 9.620923296313234e-07, "loss": 0.0101, "num_tokens": 613992597.0, "reward": 0.03515625, "reward_std": 0.05385598540306091, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 641 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.83984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1947.0, "completions/mean_length": 1848.486328125, "completions/mean_terminated_length": 802.2560424804688, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.2191687291968934, "frac_reward_zero_std": 0.9375, "grad_norm": 1.1999419115278684, "learning_rate": 9.618766697302835e-07, "loss": 0.0029, "num_tokens": 615016590.0, "reward": 0.01953125, "reward_std": 0.023823359981179237, "rewards/accuracy_reward/mean": 0.01953125, "rewards/accuracy_reward/std": 0.1385180652141571, "step": 642 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.70703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1992.0, "completions/mean_length": 1719.28125, "completions/mean_terminated_length": 925.9733276367188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.2195101135102842, "frac_reward_zero_std": 0.875, "grad_norm": 5.093660580826425, "learning_rate": 9.616604252456437e-07, "loss": -0.009, "num_tokens": 615976590.0, "reward": 0.0234375, "reward_std": 0.0479668527841568, "rewards/accuracy_reward/mean": 0.024193547666072845, "rewards/accuracy_reward/std": 0.15380480885505676, "step": 643 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.755859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2022.0, "completions/mean_length": 1792.875, "completions/mean_terminated_length": 1003.008056640625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.219851497823675, "frac_reward_zero_std": 0.75, "grad_norm": 2.8023919057887317, "learning_rate": 9.614435964843245e-07, "loss": 0.0228, "num_tokens": 616968542.0, "reward": 0.03515625, "reward_std": 0.08219823986291885, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 644 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.763671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1976.0, "completions/mean_length": 1775.263671875, "completions/mean_terminated_length": 893.942138671875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.2201928821370658, "frac_reward_zero_std": 0.875, "grad_norm": 5.44072077570558, "learning_rate": 9.612261837540738e-07, "loss": 0.0067, "num_tokens": 617952357.0, "reward": 0.033203125, "reward_std": 0.05392926186323166, "rewards/accuracy_reward/mean": 0.033203125, "rewards/accuracy_reward/std": 0.17934183776378632, "step": 645 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2026.0, "completions/mean_length": 1701.828125, "completions/mean_terminated_length": 817.1666870117188, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.2205342664504566, "frac_reward_zero_std": 0.90625, "grad_norm": 3.1143148954330266, "learning_rate": 9.610081873634696e-07, "loss": 0.0132, "num_tokens": 618912557.0, "reward": 0.015625, "reward_std": 0.03438520431518555, "rewards/accuracy_reward/mean": 0.015625, "rewards/accuracy_reward/std": 0.12414088100194931, "step": 646 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.783203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1999.0, "completions/mean_length": 1811.61328125, "completions/mean_terminated_length": 957.6396484375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.2208756507638474, "frac_reward_zero_std": 0.84375, "grad_norm": 12.358383795916899, "learning_rate": 9.607896076219181e-07, "loss": 0.0003, "num_tokens": 619921063.0, "reward": 0.06640625, "reward_std": 0.05558578670024872, "rewards/accuracy_reward/mean": 0.06640625, "rewards/accuracy_reward/std": 0.2492343932390213, "step": 647 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.728515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2021.0, "completions/mean_length": 1747.634765625, "completions/mean_terminated_length": 941.6187133789062, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.2212170350772382, "frac_reward_zero_std": 0.8125, "grad_norm": 3.6167391162865927, "learning_rate": 9.605704448396529e-07, "loss": 0.0249, "num_tokens": 620887788.0, "reward": 0.041015625, "reward_std": 0.06849661469459534, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 648 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1963.0, "completions/mean_length": 1773.365234375, "completions/mean_terminated_length": 695.9519653320312, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.221558419390629, "frac_reward_zero_std": 0.875, "grad_norm": 2.6445848095832476, "learning_rate": 9.603506993277354e-07, "loss": 0.0148, "num_tokens": 621869399.0, "reward": 0.01171875, "reward_std": 0.03697281330823898, "rewards/accuracy_reward/mean": 0.01171875, "rewards/accuracy_reward/std": 0.10772226005792618, "step": 649 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.736328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2033.0, "completions/mean_length": 1718.099609375, "completions/mean_terminated_length": 796.8222045898438, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.2218998037040198, "frac_reward_zero_std": 0.84375, "grad_norm": 5.39632894233189, "learning_rate": 9.601303713980545e-07, "loss": 0.0225, "num_tokens": 622829802.0, "reward": 0.0546875, "reward_std": 0.06546888500452042, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 650 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2032.0, "completions/mean_length": 1676.826171875, "completions/mean_terminated_length": 916.8035888671875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.2222411880174106, "frac_reward_zero_std": 0.75, "grad_norm": 5.049203049976164, "learning_rate": 9.599094613633255e-07, "loss": 0.0191, "num_tokens": 623761537.0, "reward": 0.0703125, "reward_std": 0.1024949848651886, "rewards/accuracy_reward/mean": 0.0703125, "rewards/accuracy_reward/std": 0.25592297315597534, "step": 651 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.740234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2034.0, "completions/mean_length": 1750.84375, "completions/mean_terminated_length": 904.0601806640625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.2225825723308014, "frac_reward_zero_std": 0.78125, "grad_norm": 5.696573944834077, "learning_rate": 9.596879695370894e-07, "loss": 0.013, "num_tokens": 624738305.0, "reward": 0.033203125, "reward_std": 0.07493096590042114, "rewards/accuracy_reward/mean": 0.033203125, "rewards/accuracy_reward/std": 0.17934183776378632, "step": 652 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.72265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2040.0, "completions/mean_length": 1706.591796875, "completions/mean_terminated_length": 817.0070190429688, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.2229239566441922, "frac_reward_zero_std": 0.78125, "grad_norm": 4.418678104304901, "learning_rate": 9.594658962337134e-07, "loss": 0.0226, "num_tokens": 625691632.0, "reward": 0.05859375, "reward_std": 0.07682153582572937, "rewards/accuracy_reward/mean": 0.05859375, "rewards/accuracy_reward/std": 0.23509246110916138, "step": 653 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.77734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1999.0, "completions/mean_length": 1802.8203125, "completions/mean_terminated_length": 946.8421020507812, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.223265340957583, "frac_reward_zero_std": 0.78125, "grad_norm": 8.901602624213846, "learning_rate": 9.592432417683903e-07, "loss": 0.0109, "num_tokens": 626696276.0, "reward": 0.044921875, "reward_std": 0.07729348540306091, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 654 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.77734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2023.0, "completions/mean_length": 1764.595703125, "completions/mean_terminated_length": 775.1666870117188, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.2236067252709738, "frac_reward_zero_std": 0.84375, "grad_norm": 13.61631701851183, "learning_rate": 9.59020006457137e-07, "loss": 0.0052, "num_tokens": 627672325.0, "reward": 0.05859375, "reward_std": 0.06162214279174805, "rewards/accuracy_reward/mean": 0.05859375, "rewards/accuracy_reward/std": 0.23509246110916138, "step": 655 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.736328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2043.0, "completions/mean_length": 1732.146484375, "completions/mean_terminated_length": 850.0962524414062, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.2239481095843646, "frac_reward_zero_std": 0.71875, "grad_norm": 2.0359898782231824, "learning_rate": 9.587961906167952e-07, "loss": 0.0466, "num_tokens": 628634848.0, "reward": 0.076171875, "reward_std": 0.09633205831050873, "rewards/accuracy_reward/mean": 0.076171875, "rewards/accuracy_reward/std": 0.26553234457969666, "step": 656 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2041.0, "completions/mean_length": 1797.7734375, "completions/mean_terminated_length": 816.1154174804688, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.2242894938977554, "frac_reward_zero_std": 0.8125, "grad_norm": 3.9369401285106584, "learning_rate": 9.585717945650307e-07, "loss": 0.0177, "num_tokens": 629627708.0, "reward": 0.05078125, "reward_std": 0.0761418342590332, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 657 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1818.654296875, "completions/mean_terminated_length": 960.7315063476562, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.2246308782111462, "frac_reward_zero_std": 0.78125, "grad_norm": 5.019428379206984, "learning_rate": 9.583468186203326e-07, "loss": 0.0176, "num_tokens": 630627867.0, "reward": 0.025390625, "reward_std": 0.06849660724401474, "rewards/accuracy_reward/mean": 0.026209676638245583, "rewards/accuracy_reward/std": 0.1599196493625641, "step": 658 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.791015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1921.0, "completions/mean_length": 1825.560546875, "completions/mean_terminated_length": 983.6168212890625, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.224972262524537, "frac_reward_zero_std": 0.71875, "grad_norm": 9.781776288495715, "learning_rate": 9.581212631020132e-07, "loss": 0.0185, "num_tokens": 631644218.0, "reward": 0.05859375, "reward_std": 0.09671889245510101, "rewards/accuracy_reward/mean": 0.05859375, "rewards/accuracy_reward/std": 0.23509246110916138, "step": 659 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.810546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1922.0, "completions/mean_length": 1821.67578125, "completions/mean_terminated_length": 853.3814086914062, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.2253136468379278, "frac_reward_zero_std": 0.75, "grad_norm": 6.8209333369744485, "learning_rate": 9.578951283302072e-07, "loss": 0.0074, "num_tokens": 632650564.0, "reward": 0.037109375, "reward_std": 0.08537881821393967, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 660 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.794921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1998.0, "completions/mean_length": 1823.76953125, "completions/mean_terminated_length": 954.6095581054688, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.2256550311513186, "frac_reward_zero_std": 0.78125, "grad_norm": 4.765844309007517, "learning_rate": 9.576684146258715e-07, "loss": 0.0057, "num_tokens": 633662398.0, "reward": 0.05859375, "reward_std": 0.0828578919172287, "rewards/accuracy_reward/mean": 0.05859375, "rewards/accuracy_reward/std": 0.23509246110916138, "step": 661 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2036.0, "completions/mean_length": 1830.109375, "completions/mean_terminated_length": 719.90478515625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.2259964154647094, "frac_reward_zero_std": 0.875, "grad_norm": 1.7631497884140344, "learning_rate": 9.574411223107849e-07, "loss": 0.0097, "num_tokens": 634677062.0, "reward": 0.021484375, "reward_std": 0.04604348540306091, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 662 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2020.0, "completions/mean_length": 1876.94921875, "completions/mean_terminated_length": 895.6578979492188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.2263377997781002, "frac_reward_zero_std": 0.875, "grad_norm": 2.07349670476617, "learning_rate": 9.572132517075472e-07, "loss": 0.0054, "num_tokens": 635722588.0, "reward": 0.0078125, "reward_std": 0.03125, "rewards/accuracy_reward/mean": 0.0078125, "rewards/accuracy_reward/std": 0.08812850713729858, "step": 663 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.818359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1831.556640625, "completions/mean_terminated_length": 856.3978271484375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.226679184091491, "frac_reward_zero_std": 0.8125, "grad_norm": 2.1084424092730654, "learning_rate": 9.56984803139579e-07, "loss": 0.0165, "num_tokens": 636733017.0, "reward": 0.029296875, "reward_std": 0.06425705552101135, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 664 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.865234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1952.0, "completions/mean_length": 1871.3203125, "completions/mean_terminated_length": 736.9855346679688, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.2270205684048818, "frac_reward_zero_std": 0.8125, "grad_norm": 3.303572483000552, "learning_rate": 9.567557769311213e-07, "loss": 0.0152, "num_tokens": 637772429.0, "reward": 0.015625, "reward_std": 0.05259781330823898, "rewards/accuracy_reward/mean": 0.015625, "rewards/accuracy_reward/std": 0.12414088100194931, "step": 665 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.865234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1877.0, "completions/mean_length": 1875.404296875, "completions/mean_terminated_length": 767.2898559570312, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.2273619527182726, "frac_reward_zero_std": 0.875, "grad_norm": 0.672283061801089, "learning_rate": 9.565261734072346e-07, "loss": 0.0178, "num_tokens": 638811772.0, "reward": 0.03515625, "reward_std": 0.04742163047194481, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 666 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1831.0, "completions/mean_length": 1859.57421875, "completions/mean_terminated_length": 778.6052856445312, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.2277033370316634, "frac_reward_zero_std": 0.84375, "grad_norm": 9.229973042212883, "learning_rate": 9.562959928937999e-07, "loss": 0.0061, "num_tokens": 639842706.0, "reward": 0.046875, "reward_std": 0.06304663419723511, "rewards/accuracy_reward/mean": 0.046875, "rewards/accuracy_reward/std": 0.21157780289649963, "step": 667 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1923.0, "completions/mean_length": 1880.767578125, "completions/mean_terminated_length": 620.9500122070312, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.22804472134505419, "frac_reward_zero_std": 0.9375, "grad_norm": 3.147531335995282, "learning_rate": 9.560652357175157e-07, "loss": 0.0072, "num_tokens": 640878971.0, "reward": 0.013671875, "reward_std": 0.0234375, "rewards/accuracy_reward/mean": 0.013671875, "rewards/accuracy_reward/std": 0.1162383034825325, "step": 668 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1945.0, "completions/mean_length": 1902.29296875, "completions/mean_terminated_length": 804.6333618164062, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.228386105658445, "frac_reward_zero_std": 0.875, "grad_norm": 5.107994333623763, "learning_rate": 9.558339022058995e-07, "loss": 0.0173, "num_tokens": 641930625.0, "reward": 0.02734375, "reward_std": 0.05072315037250519, "rewards/accuracy_reward/mean": 0.02822580561041832, "rewards/accuracy_reward/std": 0.1657845675945282, "step": 669 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.9296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1631.0, "completions/mean_length": 1943.947265625, "completions/mean_terminated_length": 568.138916015625, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.22872748997183578, "frac_reward_zero_std": 0.875, "grad_norm": 4.249754160057081, "learning_rate": 9.556019926872874e-07, "loss": 0.0155, "num_tokens": 643007782.0, "reward": 0.0078125, "reward_std": 0.03125, "rewards/accuracy_reward/mean": 0.0078125, "rewards/accuracy_reward/std": 0.08812850713729858, "step": 670 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.91015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1678.0, "completions/mean_length": 1931.298828125, "completions/mean_terminated_length": 749.0652465820312, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.2290688742852266, "frac_reward_zero_std": 0.75, "grad_norm": 15.119769835457573, "learning_rate": 9.553695074908321e-07, "loss": 0.015, "num_tokens": 644064607.0, "reward": 0.0390625, "reward_std": 0.08604402095079422, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 671 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.896484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1978.0, "completions/mean_length": 1933.623046875, "completions/mean_terminated_length": 943.79248046875, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.22941025859861738, "frac_reward_zero_std": 0.96875, "grad_norm": 5.792601290112887, "learning_rate": 9.55136446946504e-07, "loss": 0.0044, "num_tokens": 645132494.0, "reward": 0.001953125, "reward_std": 0.0078125, "rewards/accuracy_reward/mean": 0.001953125, "rewards/accuracy_reward/std": 0.04419417306780815, "step": 672 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2042.0, "completions/mean_length": 1920.080078125, "completions/mean_terminated_length": 879.357177734375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.2297516429120082, "frac_reward_zero_std": 0.875, "grad_norm": 2.1010597011537926, "learning_rate": 9.549028113850903e-07, "loss": 0.0082, "num_tokens": 646193399.0, "reward": 0.009765625, "reward_std": 0.03411140665411949, "rewards/accuracy_reward/mean": 0.009765625, "rewards/accuracy_reward/std": 0.09843364357948303, "step": 673 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.900390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1994.0, "completions/mean_length": 1926.142578125, "completions/mean_terminated_length": 824.6470947265625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.23009302722539898, "frac_reward_zero_std": 0.9375, "grad_norm": 5.111521833901306, "learning_rate": 9.54668601138193e-07, "loss": 0.0169, "num_tokens": 647256416.0, "reward": 0.013671875, "reward_std": 0.025633705779910088, "rewards/accuracy_reward/mean": 0.01411290280520916, "rewards/accuracy_reward/std": 0.11807556450366974, "step": 674 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.900390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1906.0, "completions/mean_length": 1914.828125, "completions/mean_terminated_length": 712.0784301757812, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.2304344115387898, "frac_reward_zero_std": 0.90625, "grad_norm": 5.817514981275791, "learning_rate": 9.544338165382318e-07, "loss": -0.001, "num_tokens": 648317544.0, "reward": 0.005859375, "reward_std": 0.0234375, "rewards/accuracy_reward/mean": 0.005859375, "rewards/accuracy_reward/std": 0.07639661431312561, "step": 675 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.90625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1944.0, "completions/mean_length": 1944.029296875, "completions/mean_terminated_length": 940.0, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.23077579585218058, "frac_reward_zero_std": 0.875, "grad_norm": 3.19701188209496, "learning_rate": 9.541984579184399e-07, "loss": 0.0055, "num_tokens": 649391975.0, "reward": 0.0078125, "reward_std": 0.03125, "rewards/accuracy_reward/mean": 0.0078125, "rewards/accuracy_reward/std": 0.08812850713729858, "step": 676 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.865234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1955.0, "completions/mean_length": 1874.87109375, "completions/mean_terminated_length": 763.3333740234375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.2311171801655714, "frac_reward_zero_std": 0.84375, "grad_norm": 21.240587694529683, "learning_rate": 9.539625256128658e-07, "loss": 0.011, "num_tokens": 650435413.0, "reward": 0.01953125, "reward_std": 0.051493462175130844, "rewards/accuracy_reward/mean": 0.01953125, "rewards/accuracy_reward/std": 0.1385180652141571, "step": 677 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.93359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1627.0, "completions/mean_length": 1958.833984375, "completions/mean_terminated_length": 705.2647094726562, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.23145856447896218, "frac_reward_zero_std": 0.90625, "grad_norm": 18.57208646125963, "learning_rate": 9.537260199563723e-07, "loss": 0.0092, "num_tokens": 651509984.0, "reward": 0.013671875, "reward_std": 0.032461829483509064, "rewards/accuracy_reward/mean": 0.013671875, "rewards/accuracy_reward/std": 0.1162383034825325, "step": 678 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1962.0, "completions/mean_length": 1914.94921875, "completions/mean_terminated_length": 912.6333618164062, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.231799948792353, "frac_reward_zero_std": 0.875, "grad_norm": 6.169590602862235, "learning_rate": 9.534889412846361e-07, "loss": 0.0114, "num_tokens": 652569190.0, "reward": 0.025390625, "reward_std": 0.04230976849794388, "rewards/accuracy_reward/mean": 0.025390625, "rewards/accuracy_reward/std": 0.15746226906776428, "step": 679 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.90234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2030.0, "completions/mean_length": 1952.755859375, "completions/mean_terminated_length": 1073.5399169921875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.23214133310574378, "frac_reward_zero_std": 0.90625, "grad_norm": 4.807154451297801, "learning_rate": 9.532512899341467e-07, "loss": 0.0134, "num_tokens": 653654393.0, "reward": 0.021484375, "reward_std": 0.03889618441462517, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 680 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.90625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1900.0, "completions/mean_length": 1925.556640625, "completions/mean_terminated_length": 741.9375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.2324827174191346, "frac_reward_zero_std": 0.9375, "grad_norm": 1.9076291352445545, "learning_rate": 9.53013066242207e-07, "loss": 0.0049, "num_tokens": 654714166.0, "reward": 0.00390625, "reward_std": 0.015625, "rewards/accuracy_reward/mean": 0.00390625, "rewards/accuracy_reward/std": 0.06243881583213806, "step": 681 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.92578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1753.0, "completions/mean_length": 1947.02734375, "completions/mean_terminated_length": 687.5263061523438, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.23282410173252538, "frac_reward_zero_std": 0.84375, "grad_norm": 9.421055299085438, "learning_rate": 9.527742705469318e-07, "loss": 0.0214, "num_tokens": 655786836.0, "reward": 0.02734375, "reward_std": 0.0580955371260643, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 682 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.9140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1979.0, "completions/mean_length": 1931.18359375, "completions/mean_terminated_length": 689.9318237304688, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.2331654860459162, "frac_reward_zero_std": 0.90625, "grad_norm": 5.449464217623047, "learning_rate": 9.525349031872481e-07, "loss": 0.0136, "num_tokens": 656852226.0, "reward": 0.013671875, "reward_std": 0.032461829483509064, "rewards/accuracy_reward/mean": 0.013671875, "rewards/accuracy_reward/std": 0.1162383034825325, "step": 683 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.93359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2033.0, "completions/mean_length": 1964.025390625, "completions/mean_terminated_length": 783.441162109375, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.23350687035930698, "frac_reward_zero_std": 0.875, "grad_norm": 19.298173662881958, "learning_rate": 9.52294964502894e-07, "loss": 0.0173, "num_tokens": 657932031.0, "reward": 0.01953125, "reward_std": 0.046542368829250336, "rewards/accuracy_reward/mean": 0.01953125, "rewards/accuracy_reward/std": 0.1385180652141571, "step": 684 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.87109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2005.0, "completions/mean_length": 1891.552734375, "completions/mean_terminated_length": 834.3485107421875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.2338482546726978, "frac_reward_zero_std": 0.875, "grad_norm": 7.337417659861314, "learning_rate": 9.520544548344184e-07, "loss": -0.0011, "num_tokens": 658979626.0, "reward": 0.01171875, "reward_std": 0.03697281330823898, "rewards/accuracy_reward/mean": 0.01171875, "rewards/accuracy_reward/std": 0.10772226005792618, "step": 685 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.900390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1731.0, "completions/mean_length": 1910.216796875, "completions/mean_terminated_length": 664.7647094726562, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.23418963898608858, "frac_reward_zero_std": 0.875, "grad_norm": 7.425422756588813, "learning_rate": 9.51813374523181e-07, "loss": 0.0066, "num_tokens": 660033913.0, "reward": 0.01171875, "reward_std": 0.03697281330823898, "rewards/accuracy_reward/mean": 0.01171875, "rewards/accuracy_reward/std": 0.10772226005792618, "step": 686 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2018.0, "completions/mean_length": 1895.041015625, "completions/mean_terminated_length": 824.328125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.2345310232994794, "frac_reward_zero_std": 0.84375, "grad_norm": 8.225695329815126, "learning_rate": 9.515717239113511e-07, "loss": 0.0094, "num_tokens": 661083550.0, "reward": 0.041015625, "reward_std": 0.06790468841791153, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 687 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.857421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1969.0, "completions/mean_length": 1883.64453125, "completions/mean_terminated_length": 895.9862670898438, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.23487240761287018, "frac_reward_zero_std": 0.84375, "grad_norm": 4.550178195000361, "learning_rate": 9.513295033419077e-07, "loss": 0.0085, "num_tokens": 662136680.0, "reward": 0.029296875, "reward_std": 0.05628519505262375, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 688 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.84765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1989.0, "completions/mean_length": 1857.345703125, "completions/mean_terminated_length": 796.525634765625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.235213791926261, "frac_reward_zero_std": 0.9375, "grad_norm": 0.8550177090278439, "learning_rate": 9.510867131586383e-07, "loss": 0.0112, "num_tokens": 663167625.0, "reward": 0.00390625, "reward_std": 0.015625, "rewards/accuracy_reward/mean": 0.004032257944345474, "rewards/accuracy_reward/std": 0.06343588978052139, "step": 689 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.822265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1950.0, "completions/mean_length": 1828.4375, "completions/mean_terminated_length": 812.6593627929688, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.23555517623965178, "frac_reward_zero_std": 0.90625, "grad_norm": 9.326333054201168, "learning_rate": 9.508433537061394e-07, "loss": -0.0012, "num_tokens": 664183097.0, "reward": 0.009765625, "reward_std": 0.028222277760505676, "rewards/accuracy_reward/mean": 0.010080644860863686, "rewards/accuracy_reward/std": 0.0999959334731102, "step": 690 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.884765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1705.0, "completions/mean_length": 1912.939453125, "completions/mean_terminated_length": 875.9491577148438, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.2358965605530426, "frac_reward_zero_std": 0.875, "grad_norm": 8.407769084194278, "learning_rate": 9.505994253298152e-07, "loss": 0.0059, "num_tokens": 665239114.0, "reward": 0.01171875, "reward_std": 0.036034777760505676, "rewards/accuracy_reward/mean": 0.01171875, "rewards/accuracy_reward/std": 0.10772226005792618, "step": 691 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1984.0, "completions/mean_length": 1845.53515625, "completions/mean_terminated_length": 921.7391357421875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.23623794486643337, "frac_reward_zero_std": 0.9375, "grad_norm": 5.225003508802906, "learning_rate": 9.503549283758773e-07, "loss": 0.0021, "num_tokens": 666264828.0, "reward": 0.021484375, "reward_std": 0.02998628467321396, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 692 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.837890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1981.0, "completions/mean_length": 1840.1484375, "completions/mean_terminated_length": 765.831298828125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.2365793291798242, "frac_reward_zero_std": 0.875, "grad_norm": 3.442083675696622, "learning_rate": 9.501098631913446e-07, "loss": 0.0123, "num_tokens": 667276664.0, "reward": 0.0234375, "reward_std": 0.03944835811853409, "rewards/accuracy_reward/mean": 0.0234375, "rewards/accuracy_reward/std": 0.15143637359142303, "step": 693 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.884765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2012.0, "completions/mean_length": 1911.650390625, "completions/mean_terminated_length": 864.7626953125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.23692071349321497, "frac_reward_zero_std": 0.9375, "grad_norm": 7.307947045544329, "learning_rate": 9.498642301240422e-07, "loss": 0.0045, "num_tokens": 668327941.0, "reward": 0.013671875, "reward_std": 0.025633705779910088, "rewards/accuracy_reward/mean": 0.013671875, "rewards/accuracy_reward/std": 0.1162383034825325, "step": 694 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.900390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2027.0, "completions/mean_length": 1922.025390625, "completions/mean_terminated_length": 783.3137817382812, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.2372620978066058, "frac_reward_zero_std": 0.875, "grad_norm": 1.0203181571697983, "learning_rate": 9.496180295226012e-07, "loss": 0.0248, "num_tokens": 669385618.0, "reward": 0.01171875, "reward_std": 0.036034777760505676, "rewards/accuracy_reward/mean": 0.01171875, "rewards/accuracy_reward/std": 0.10772226005792618, "step": 695 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1934.0, "completions/mean_length": 1844.013671875, "completions/mean_terminated_length": 960.0729370117188, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.23760348211999657, "frac_reward_zero_std": 0.84375, "grad_norm": 13.45281910422109, "learning_rate": 9.493712617364585e-07, "loss": 0.0207, "num_tokens": 670409193.0, "reward": 0.037109375, "reward_std": 0.058760739862918854, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 696 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.861328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2013.0, "completions/mean_length": 1870.7890625, "completions/mean_terminated_length": 770.08447265625, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.2379448664333874, "frac_reward_zero_std": 0.84375, "grad_norm": 7.606653637945903, "learning_rate": 9.491239271158558e-07, "loss": -0.0012, "num_tokens": 671443197.0, "reward": 0.041015625, "reward_std": 0.050508126616477966, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 697 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.85546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1974.0, "completions/mean_length": 1871.021484375, "completions/mean_terminated_length": 823.5, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.23828625074677817, "frac_reward_zero_std": 0.90625, "grad_norm": 2.834670748941362, "learning_rate": 9.488760260118393e-07, "loss": 0.0025, "num_tokens": 672475544.0, "reward": 0.01953125, "reward_std": 0.03697281330823898, "rewards/accuracy_reward/mean": 0.01953125, "rewards/accuracy_reward/std": 0.1385180652141571, "step": 698 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.779296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1847.0, "completions/mean_length": 1777.111328125, "completions/mean_terminated_length": 820.610595703125, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.23862763506016899, "frac_reward_zero_std": 0.78125, "grad_norm": 15.00588616683939, "learning_rate": 9.486275587762592e-07, "loss": 0.0042, "num_tokens": 673457585.0, "reward": 0.037109375, "reward_std": 0.07917051017284393, "rewards/accuracy_reward/mean": 0.038306452333927155, "rewards/accuracy_reward/std": 0.19212883710861206, "step": 699 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2034.0, "completions/mean_length": 1821.599609375, "completions/mean_terminated_length": 730.7614135742188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.23896901937355977, "frac_reward_zero_std": 0.75, "grad_norm": 11.245960679230839, "learning_rate": 9.483785257617695e-07, "loss": 0.005, "num_tokens": 674460884.0, "reward": 0.06640625, "reward_std": 0.09743183851242065, "rewards/accuracy_reward/mean": 0.06640625, "rewards/accuracy_reward/std": 0.2492343932390213, "step": 700 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.849609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1915.0, "completions/mean_length": 1849.265625, "completions/mean_terminated_length": 726.5454711914062, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.23931040368695058, "frac_reward_zero_std": 0.875, "grad_norm": 14.160772576169151, "learning_rate": 9.48128927321827e-07, "loss": -0.0006, "num_tokens": 675486732.0, "reward": 0.05078125, "reward_std": 0.061169594526290894, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 701 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.865234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1860.0, "completions/mean_length": 1860.3203125, "completions/mean_terminated_length": 655.3623046875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.2396517880003414, "frac_reward_zero_std": 0.90625, "grad_norm": 2.898876186601683, "learning_rate": 9.478787638106908e-07, "loss": 0.0099, "num_tokens": 676522128.0, "reward": 0.0390625, "reward_std": 0.04243633896112442, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 702 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1962.0, "completions/mean_length": 1909.7421875, "completions/mean_terminated_length": 686.6923217773438, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.23999317231373218, "frac_reward_zero_std": 0.90625, "grad_norm": 3.3994186384542977, "learning_rate": 9.476280355834224e-07, "loss": 0.0041, "num_tokens": 677568652.0, "reward": 0.017578125, "reward_std": 0.03536957502365112, "rewards/accuracy_reward/mean": 0.017578125, "rewards/accuracy_reward/std": 0.13154059648513794, "step": 703 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2006.0, "completions/mean_length": 1921.705078125, "completions/mean_terminated_length": 804.4807739257812, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.240334556627123, "frac_reward_zero_std": 0.9375, "grad_norm": 1.0858432771960347, "learning_rate": 9.473767429958846e-07, "loss": 0.0081, "num_tokens": 678628245.0, "reward": 0.009765625, "reward_std": 0.023271184414625168, "rewards/accuracy_reward/mean": 0.009765625, "rewards/accuracy_reward/std": 0.09843364357948303, "step": 704 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.814453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2019.0, "completions/mean_length": 1821.080078125, "completions/mean_terminated_length": 825.431640625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.24067594094051378, "frac_reward_zero_std": 0.6875, "grad_norm": 11.609559062075409, "learning_rate": 9.471248864047415e-07, "loss": 0.0312, "num_tokens": 679635278.0, "reward": 0.0859375, "reward_std": 0.11311560124158859, "rewards/accuracy_reward/mean": 0.08870967477560043, "rewards/accuracy_reward/std": 0.2846112847328186, "step": 705 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.830078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2018.0, "completions/mean_length": 1832.087890625, "completions/mean_terminated_length": 777.3448486328125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.2410173252539046, "frac_reward_zero_std": 0.875, "grad_norm": 8.27326656183113, "learning_rate": 9.468724661674571e-07, "loss": 0.0126, "num_tokens": 680654187.0, "reward": 0.041015625, "reward_std": 0.046722229570150375, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 706 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.83203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1885.0, "completions/mean_length": 1812.83203125, "completions/mean_terminated_length": 647.9302368164062, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.24135870956729538, "frac_reward_zero_std": 0.875, "grad_norm": 5.026506072673076, "learning_rate": 9.466194826422961e-07, "loss": 0.006, "num_tokens": 681664693.0, "reward": 0.021484375, "reward_std": 0.04698248207569122, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 707 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1884.0, "completions/mean_length": 1820.4765625, "completions/mean_terminated_length": 834.5416870117188, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.2417000938806862, "frac_reward_zero_std": 0.84375, "grad_norm": 15.492257018422846, "learning_rate": 9.463659361883219e-07, "loss": 0.0052, "num_tokens": 682677961.0, "reward": 0.015625, "reward_std": 0.047646719962358475, "rewards/accuracy_reward/mean": 0.015625, "rewards/accuracy_reward/std": 0.12414088100194931, "step": 708 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1926.0, "completions/mean_length": 1926.41796875, "completions/mean_terminated_length": 936.3928833007812, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.24204147819407698, "frac_reward_zero_std": 0.90625, "grad_norm": 2.369205753473792, "learning_rate": 9.46111827165398e-07, "loss": 0.0032, "num_tokens": 683733519.0, "reward": 0.015625, "reward_std": 0.03438520431518555, "rewards/accuracy_reward/mean": 0.015625, "rewards/accuracy_reward/std": 0.12414088100194931, "step": 709 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.80078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2021.0, "completions/mean_length": 1772.890625, "completions/mean_terminated_length": 667.058837890625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.2423828625074678, "frac_reward_zero_std": 0.90625, "grad_norm": 4.6176891808664795, "learning_rate": 9.458571559341849e-07, "loss": 0.0073, "num_tokens": 684712423.0, "reward": 0.04296875, "reward_std": 0.042583562433719635, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 710 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2036.0, "completions/mean_length": 1816.73046875, "completions/mean_terminated_length": 638.357177734375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.24272424682085858, "frac_reward_zero_std": 0.78125, "grad_norm": 4.10296319651108, "learning_rate": 9.456019228561425e-07, "loss": 0.0281, "num_tokens": 685719165.0, "reward": 0.0625, "reward_std": 0.0932510495185852, "rewards/accuracy_reward/mean": 0.06451612710952759, "rewards/accuracy_reward/std": 0.2459181249141693, "step": 711 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1974.0, "completions/mean_length": 1898.576171875, "completions/mean_terminated_length": 922.9264526367188, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.2430656311342494, "frac_reward_zero_std": 0.875, "grad_norm": 3.4854106119832857, "learning_rate": 9.453461282935271e-07, "loss": 0.0063, "num_tokens": 686766788.0, "reward": 0.02734375, "reward_std": 0.04517117142677307, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 712 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.880859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1771.0, "completions/mean_length": 1903.732421875, "completions/mean_terminated_length": 837.0983276367188, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.24340701544764018, "frac_reward_zero_std": 0.9375, "grad_norm": 1.7146508048265992, "learning_rate": 9.450897726093924e-07, "loss": 0.0048, "num_tokens": 687815963.0, "reward": 0.01171875, "reward_std": 0.022772299125790596, "rewards/accuracy_reward/mean": 0.01171875, "rewards/accuracy_reward/std": 0.10772226005792618, "step": 713 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.876953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1893.0, "completions/mean_length": 1898.29296875, "completions/mean_terminated_length": 831.3333740234375, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.243748399761031, "frac_reward_zero_std": 0.90625, "grad_norm": 5.274635279469366, "learning_rate": 9.448328561675883e-07, "loss": -0.0013, "num_tokens": 688864945.0, "reward": 0.029296875, "reward_std": 0.03741292655467987, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 714 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1889.0, "completions/mean_length": 1948.87109375, "completions/mean_terminated_length": 779.1500244140625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.24408978407442178, "frac_reward_zero_std": 0.90625, "grad_norm": 1.0808293627719798, "learning_rate": 9.44575379332761e-07, "loss": 0.0076, "num_tokens": 689933231.0, "reward": 0.0390625, "reward_std": 0.02629890665411949, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 715 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.935546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2040.0, "completions/mean_length": 1976.615234375, "completions/mean_terminated_length": 940.45458984375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.2444311683878126, "frac_reward_zero_std": 0.96875, "grad_norm": 3.257313844473362, "learning_rate": 9.443173424703514e-07, "loss": 0.0026, "num_tokens": 691023322.0, "reward": 0.00390625, "reward_std": 0.010673906654119492, "rewards/accuracy_reward/mean": 0.004032257944345474, "rewards/accuracy_reward/std": 0.06343588978052139, "step": 716 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1964.0, "completions/mean_length": 1841.89453125, "completions/mean_terminated_length": 848.8409423828125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.24477255270120338, "frac_reward_zero_std": 0.84375, "grad_norm": 12.556183918575925, "learning_rate": 9.440587459465956e-07, "loss": 0.0085, "num_tokens": 692040212.0, "reward": 0.021484375, "reward_std": 0.054354868829250336, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 717 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.884765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2007.0, "completions/mean_length": 1895.978515625, "completions/mean_terminated_length": 728.7626953125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.2451139370145942, "frac_reward_zero_std": 0.84375, "grad_norm": 8.641470017684496, "learning_rate": 9.437995901285246e-07, "loss": 0.0063, "num_tokens": 693085721.0, "reward": 0.033203125, "reward_std": 0.05793476849794388, "rewards/accuracy_reward/mean": 0.033203125, "rewards/accuracy_reward/std": 0.17934183776378632, "step": 718 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1747.0, "completions/mean_length": 1903.697265625, "completions/mean_terminated_length": 728.6607666015625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.24545532132798498, "frac_reward_zero_std": 0.84375, "grad_norm": 17.175972875074105, "learning_rate": 9.435398753839622e-07, "loss": 0.0055, "num_tokens": 694137118.0, "reward": 0.01171875, "reward_std": 0.04192390665411949, "rewards/accuracy_reward/mean": 0.01171875, "rewards/accuracy_reward/std": 0.10772226005792618, "step": 719 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.9140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1774.0, "completions/mean_length": 1908.376953125, "completions/mean_terminated_length": 423.29547119140625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.2457967056413758, "frac_reward_zero_std": 1.0, "grad_norm": 0.0, "learning_rate": 9.432796020815261e-07, "loss": 0.0, "num_tokens": 695184159.0, "reward": 0.0, "reward_std": 0.0, "rewards/accuracy_reward/mean": 0.0, "rewards/accuracy_reward/std": 0.0, "step": 720 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.89453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2032.0, "completions/mean_length": 1900.154296875, "completions/mean_terminated_length": 646.2037353515625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.24613808995476658, "frac_reward_zero_std": 0.90625, "grad_norm": 6.07614082677678, "learning_rate": 9.430187705906268e-07, "loss": -0.005, "num_tokens": 696227726.0, "reward": 0.01171875, "reward_std": 0.02960042469203472, "rewards/accuracy_reward/mean": 0.01171875, "rewards/accuracy_reward/std": 0.10772226005792618, "step": 721 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.87109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1878.0, "completions/mean_length": 1888.22265625, "completions/mean_terminated_length": 808.5151977539062, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.2464794742681574, "frac_reward_zero_std": 0.875, "grad_norm": 10.735065726966466, "learning_rate": 9.427573812814666e-07, "loss": 0.0016, "num_tokens": 697261488.0, "reward": 0.01171875, "reward_std": 0.03697281330823898, "rewards/accuracy_reward/mean": 0.01171875, "rewards/accuracy_reward/std": 0.10772226005792618, "step": 722 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1952.0, "completions/mean_length": 1955.484375, "completions/mean_terminated_length": 863.7999877929688, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.24682085858154817, "frac_reward_zero_std": 0.90625, "grad_norm": 33.91462387898152, "learning_rate": 9.424954345250401e-07, "loss": 0.0015, "num_tokens": 698336456.0, "reward": 0.005859375, "reward_std": 0.0234375, "rewards/accuracy_reward/mean": 0.005859375, "rewards/accuracy_reward/std": 0.07639661431312561, "step": 723 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1912.0, "completions/mean_length": 1828.763671875, "completions/mean_terminated_length": 711.702392578125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.247162242894939, "frac_reward_zero_std": 0.84375, "grad_norm": 6.5097438308450295, "learning_rate": 9.422329306931325e-07, "loss": 0.0029, "num_tokens": 699354207.0, "reward": 0.01171875, "reward_std": 0.04192390665411949, "rewards/accuracy_reward/mean": 0.01171875, "rewards/accuracy_reward/std": 0.10772226005792618, "step": 724 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.888671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2024.0, "completions/mean_length": 1896.5625, "completions/mean_terminated_length": 687.7192993164062, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.24750362720832977, "frac_reward_zero_std": 0.9375, "grad_norm": 4.253917350886468, "learning_rate": 9.419698701583204e-07, "loss": 0.0047, "num_tokens": 700401263.0, "reward": 0.00390625, "reward_std": 0.015625, "rewards/accuracy_reward/mean": 0.004032257944345474, "rewards/accuracy_reward/std": 0.06343588978052139, "step": 725 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.873046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1819.0, "completions/mean_length": 1875.80078125, "completions/mean_terminated_length": 691.5999755859375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.2478450115217206, "frac_reward_zero_std": 0.96875, "grad_norm": 3.266792260067484, "learning_rate": 9.417062532939698e-07, "loss": -0.0013, "num_tokens": 701444025.0, "reward": 0.001953125, "reward_std": 0.0078125, "rewards/accuracy_reward/mean": 0.001953125, "rewards/accuracy_reward/std": 0.04419417306780815, "step": 726 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.818359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2039.0, "completions/mean_length": 1858.53515625, "completions/mean_terminated_length": 1005.3118286132812, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.24818639583511137, "frac_reward_zero_std": 0.90625, "grad_norm": 6.13482927774544, "learning_rate": 9.414420804742366e-07, "loss": 0.0065, "num_tokens": 702477787.0, "reward": 0.015625, "reward_std": 0.03438520431518555, "rewards/accuracy_reward/mean": 0.015625, "rewards/accuracy_reward/std": 0.12414088100194931, "step": 727 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.865234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2004.0, "completions/mean_length": 1878.255859375, "completions/mean_terminated_length": 788.4492797851562, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.2485277801485022, "frac_reward_zero_std": 0.90625, "grad_norm": 5.75132103856005, "learning_rate": 9.41177352074066e-07, "loss": 0.0069, "num_tokens": 703516622.0, "reward": 0.017578125, "reward_std": 0.03576334938406944, "rewards/accuracy_reward/mean": 0.018145160749554634, "rewards/accuracy_reward/std": 0.1336110234260559, "step": 728 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.9296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1916.0, "completions/mean_length": 1966.896484375, "completions/mean_terminated_length": 894.5277709960938, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.24886916446189297, "frac_reward_zero_std": 0.84375, "grad_norm": 11.543322958878461, "learning_rate": 9.409120684691915e-07, "loss": 0.0023, "num_tokens": 704599193.0, "reward": 0.03515625, "reward_std": 0.05971328169107437, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 729 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.87109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2028.0, "completions/mean_length": 1879.03125, "completions/mean_terminated_length": 737.212158203125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.24921054877528379, "frac_reward_zero_std": 0.8125, "grad_norm": 9.378120638546573, "learning_rate": 9.406462300361345e-07, "loss": 0.0222, "num_tokens": 705633241.0, "reward": 0.03515625, "reward_std": 0.062220655381679535, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 730 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.849609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1921.0, "completions/mean_length": 1870.765625, "completions/mean_terminated_length": 869.5064697265625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.24955193308867457, "frac_reward_zero_std": 0.90625, "grad_norm": 1.8710887732685961, "learning_rate": 9.403798371522042e-07, "loss": 0.0019, "num_tokens": 706669809.0, "reward": 0.01171875, "reward_std": 0.02960042469203472, "rewards/accuracy_reward/mean": 0.01171875, "rewards/accuracy_reward/std": 0.10772226005792618, "step": 731 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.916015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1700.0, "completions/mean_length": 1922.0078125, "completions/mean_terminated_length": 547.81396484375, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.24989331740206538, "frac_reward_zero_std": 0.90625, "grad_norm": 4.308225203261697, "learning_rate": 9.401128901954964e-07, "loss": 0.012, "num_tokens": 707725973.0, "reward": 0.013671875, "reward_std": 0.032461829483509064, "rewards/accuracy_reward/mean": 0.013671875, "rewards/accuracy_reward/std": 0.1162383034825325, "step": 732 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.794921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2030.0, "completions/mean_length": 1777.404296875, "completions/mean_terminated_length": 728.5238647460938, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.2502347017154562, "frac_reward_zero_std": 0.71875, "grad_norm": 6.096172823484754, "learning_rate": 9.398453895448936e-07, "loss": 0.028, "num_tokens": 708721348.0, "reward": 0.033203125, "reward_std": 0.08890639245510101, "rewards/accuracy_reward/mean": 0.033203125, "rewards/accuracy_reward/std": 0.17934183776378632, "step": 733 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.736328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2012.0, "completions/mean_length": 1750.755859375, "completions/mean_terminated_length": 920.674072265625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.25057608602884696, "frac_reward_zero_std": 0.84375, "grad_norm": 6.2339362130962455, "learning_rate": 9.395773355800643e-07, "loss": 0.0022, "num_tokens": 709699703.0, "reward": 0.03125, "reward_std": 0.05789502337574959, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 734 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2036.0, "completions/mean_length": 1846.318359375, "completions/mean_terminated_length": 874.57958984375, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.25091747034223777, "frac_reward_zero_std": 0.8125, "grad_norm": 4.351020923091093, "learning_rate": 9.393087286814616e-07, "loss": 0.0235, "num_tokens": 710723146.0, "reward": 0.037109375, "reward_std": 0.06860867142677307, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 735 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.849609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1733.0, "completions/mean_length": 1859.740234375, "completions/mean_terminated_length": 796.19482421875, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.2512588546556286, "frac_reward_zero_std": 0.8125, "grad_norm": 11.218381978402311, "learning_rate": 9.39039569230324e-07, "loss": 0.0127, "num_tokens": 711753461.0, "reward": 0.037109375, "reward_std": 0.0723423883318901, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 736 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.787109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2006.0, "completions/mean_length": 1780.439453125, "completions/mean_terminated_length": 791.2017822265625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.2516002389690194, "frac_reward_zero_std": 0.84375, "grad_norm": 8.405206048368234, "learning_rate": 9.387698576086743e-07, "loss": -0.0006, "num_tokens": 712735094.0, "reward": 0.068359375, "reward_std": 0.06536377221345901, "rewards/accuracy_reward/mean": 0.07056451588869095, "rewards/accuracy_reward/std": 0.25635457038879395, "step": 737 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.728515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1987.0, "completions/mean_length": 1702.953125, "completions/mean_terminated_length": 777.0360107421875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.25194162328241015, "frac_reward_zero_std": 0.71875, "grad_norm": 8.361406843910276, "learning_rate": 9.384995941993187e-07, "loss": 0.0029, "num_tokens": 713679950.0, "reward": 0.05859375, "reward_std": 0.09908141195774078, "rewards/accuracy_reward/mean": 0.05859375, "rewards/accuracy_reward/std": 0.23509246110916138, "step": 738 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.75, "completions/max_length": 2048.0, "completions/max_terminated_length": 1861.0, "completions/mean_length": 1730.751953125, "completions/mean_terminated_length": 779.0078125, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.25228300759580097, "frac_reward_zero_std": 0.75, "grad_norm": 5.992097237969598, "learning_rate": 9.382287793858467e-07, "loss": 0.0164, "num_tokens": 714641135.0, "reward": 0.064453125, "reward_std": 0.10009271651506424, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 739 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1911.0, "completions/mean_length": 1734.02734375, "completions/mean_terminated_length": 830.1666870117188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.2526243919091918, "frac_reward_zero_std": 0.78125, "grad_norm": 13.872577742999004, "learning_rate": 9.379574135526304e-07, "loss": 0.0302, "num_tokens": 715606413.0, "reward": 0.044921875, "reward_std": 0.08235900849103928, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 740 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.787109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1989.0, "completions/mean_length": 1817.380859375, "completions/mean_terminated_length": 964.7247314453125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.2529657762225826, "frac_reward_zero_std": 0.84375, "grad_norm": 17.771672883756676, "learning_rate": 9.376854970848239e-07, "loss": 0.0094, "num_tokens": 716612304.0, "reward": 0.03515625, "reward_std": 0.05001020431518555, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 741 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.76171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2018.0, "completions/mean_length": 1753.06640625, "completions/mean_terminated_length": 810.245849609375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.25330716053597335, "frac_reward_zero_std": 0.90625, "grad_norm": 6.652605277169814, "learning_rate": 9.374130303683628e-07, "loss": 0.0128, "num_tokens": 717594754.0, "reward": 0.021484375, "reward_std": 0.03735867142677307, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 742 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.701171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1956.0, "completions/mean_length": 1720.76171875, "completions/mean_terminated_length": 952.9281005859375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.25364854484936417, "frac_reward_zero_std": 0.875, "grad_norm": 4.997219291130501, "learning_rate": 9.371400137899642e-07, "loss": 0.0029, "num_tokens": 718554424.0, "reward": 0.0390625, "reward_std": 0.05243149772286415, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 743 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.791015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1871.0, "completions/mean_length": 1821.517578125, "completions/mean_terminated_length": 964.27099609375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.253989929162755, "frac_reward_zero_std": 0.90625, "grad_norm": 1.3196305678553888, "learning_rate": 9.368664477371246e-07, "loss": 0.0014, "num_tokens": 719560033.0, "reward": 0.009765625, "reward_std": 0.029160313308238983, "rewards/accuracy_reward/mean": 0.009765625, "rewards/accuracy_reward/std": 0.09843364357948303, "step": 744 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2041.0, "completions/mean_length": 1610.068359375, "completions/mean_terminated_length": 829.4076538085938, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.2543313134761458, "frac_reward_zero_std": 0.625, "grad_norm": 4.516296029437842, "learning_rate": 9.365923325981214e-07, "loss": 0.0195, "num_tokens": 720459588.0, "reward": 0.080078125, "reward_std": 0.12215207517147064, "rewards/accuracy_reward/mean": 0.080078125, "rewards/accuracy_reward/std": 0.271679550409317, "step": 745 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.666015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2016.0, "completions/mean_length": 1637.5625, "completions/mean_terminated_length": 819.0877075195312, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.25467269778953655, "frac_reward_zero_std": 0.8125, "grad_norm": 4.358497047175415, "learning_rate": 9.363176687620109e-07, "loss": -0.0001, "num_tokens": 721375460.0, "reward": 0.02734375, "reward_std": 0.06233368441462517, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 746 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.732421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2008.0, "completions/mean_length": 1719.345703125, "completions/mean_terminated_length": 819.7445068359375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.25501408210292736, "frac_reward_zero_std": 0.84375, "grad_norm": 1.059356007926346, "learning_rate": 9.360424566186279e-07, "loss": 0.0193, "num_tokens": 722337493.0, "reward": 0.01171875, "reward_std": 0.04192390665411949, "rewards/accuracy_reward/mean": 0.01171875, "rewards/accuracy_reward/std": 0.10772226005792618, "step": 747 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.755859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2014.0, "completions/mean_length": 1793.783203125, "completions/mean_terminated_length": 1006.72802734375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.2553554664163182, "frac_reward_zero_std": 0.84375, "grad_norm": 4.417724225579728, "learning_rate": 9.357666965585859e-07, "loss": 0.0056, "num_tokens": 723329686.0, "reward": 0.05078125, "reward_std": 0.05237272381782532, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 748 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.646484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1990.0, "completions/mean_length": 1638.3203125, "completions/mean_terminated_length": 889.1271362304688, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.255696850729709, "frac_reward_zero_std": 0.84375, "grad_norm": 6.572466992615142, "learning_rate": 9.354903889732761e-07, "loss": 0.0093, "num_tokens": 724254314.0, "reward": 0.025390625, "reward_std": 0.052045635879039764, "rewards/accuracy_reward/mean": 0.025390625, "rewards/accuracy_reward/std": 0.15746226906776428, "step": 749 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.794921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1952.0, "completions/mean_length": 1785.15234375, "completions/mean_terminated_length": 766.3048095703125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.25603823504309975, "frac_reward_zero_std": 0.8125, "grad_norm": 5.952264768861366, "learning_rate": 9.352135342548659e-07, "loss": 0.0111, "num_tokens": 725243624.0, "reward": 0.037109375, "reward_std": 0.07174387574195862, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 750 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.701171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1972.0, "completions/mean_length": 1695.29296875, "completions/mean_terminated_length": 867.6993408203125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.25637961935649056, "frac_reward_zero_std": 0.78125, "grad_norm": 2.104368814767614, "learning_rate": 9.349361327963006e-07, "loss": 0.0316, "num_tokens": 726191022.0, "reward": 0.056640625, "reward_std": 0.08149883151054382, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 751 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1998.0, "completions/mean_length": 1723.990234375, "completions/mean_terminated_length": 911.74658203125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.2567210036698814, "frac_reward_zero_std": 0.875, "grad_norm": 0.9758544955350608, "learning_rate": 9.346581849913004e-07, "loss": 0.0091, "num_tokens": 727153753.0, "reward": 0.01171875, "reward_std": 0.036034777760505676, "rewards/accuracy_reward/mean": 0.01171875, "rewards/accuracy_reward/std": 0.10772226005792618, "step": 752 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.66796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1962.0, "completions/mean_length": 1654.111328125, "completions/mean_terminated_length": 861.7000122070312, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.2570623879832722, "frac_reward_zero_std": 0.78125, "grad_norm": 9.588401547947068, "learning_rate": 9.343796912343617e-07, "loss": 0.0088, "num_tokens": 728077778.0, "reward": 0.05859375, "reward_std": 0.0792168527841568, "rewards/accuracy_reward/mean": 0.05859375, "rewards/accuracy_reward/std": 0.23509246110916138, "step": 753 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2000.0, "completions/mean_length": 1764.6484375, "completions/mean_terminated_length": 878.0322265625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.25740377229666295, "frac_reward_zero_std": 0.78125, "grad_norm": 8.455847970024246, "learning_rate": 9.341006519207551e-07, "loss": 0.0106, "num_tokens": 729061630.0, "reward": 0.029296875, "reward_std": 0.07179813086986542, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 754 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2035.0, "completions/mean_length": 1661.6953125, "completions/mean_terminated_length": 870.6904907226562, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.25774515661005376, "frac_reward_zero_std": 0.78125, "grad_norm": 5.735508205961146, "learning_rate": 9.338210674465263e-07, "loss": 0.0217, "num_tokens": 729985842.0, "reward": 0.060546875, "reward_std": 0.080595001578331, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 755 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.69921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1966.0, "completions/mean_length": 1672.8671875, "completions/mean_terminated_length": 800.80517578125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.2580865409234446, "frac_reward_zero_std": 0.78125, "grad_norm": 17.12835869836486, "learning_rate": 9.335409382084939e-07, "loss": 0.0053, "num_tokens": 730918398.0, "reward": 0.056640625, "reward_std": 0.08643084019422531, "rewards/accuracy_reward/mean": 0.058467742055654526, "rewards/accuracy_reward/std": 0.23486268520355225, "step": 756 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.73046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1860.0, "completions/mean_length": 1720.318359375, "completions/mean_terminated_length": 832.253662109375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.2584279252368354, "frac_reward_zero_std": 0.71875, "grad_norm": 1.2749479704223878, "learning_rate": 9.332602646042504e-07, "loss": 0.007, "num_tokens": 731877809.0, "reward": 0.044921875, "reward_std": 0.09522771835327148, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 757 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.66796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.0, "completions/mean_length": 1669.61328125, "completions/mean_terminated_length": 908.3882446289062, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.25876930955022615, "frac_reward_zero_std": 0.8125, "grad_norm": 1.071438862279623, "learning_rate": 9.329790470321607e-07, "loss": 0.0219, "num_tokens": 732812267.0, "reward": 0.041015625, "reward_std": 0.07379381358623505, "rewards/accuracy_reward/mean": 0.04233871027827263, "rewards/accuracy_reward/std": 0.2015640139579773, "step": 758 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2037.0, "completions/mean_length": 1748.84765625, "completions/mean_terminated_length": 887.6515502929688, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.25911069386361696, "frac_reward_zero_std": 0.875, "grad_norm": 10.401037798775814, "learning_rate": 9.326972858913613e-07, "loss": 0.0065, "num_tokens": 733780477.0, "reward": 0.0390625, "reward_std": 0.058208562433719635, "rewards/accuracy_reward/mean": 0.04032257944345474, "rewards/accuracy_reward/std": 0.19691328704357147, "step": 759 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.771484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2026.0, "completions/mean_length": 1767.48046875, "completions/mean_terminated_length": 820.4273681640625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.25945207817700777, "frac_reward_zero_std": 0.90625, "grad_norm": 12.927825227835879, "learning_rate": 9.324149815817612e-07, "loss": 0.0073, "num_tokens": 734765827.0, "reward": 0.017578125, "reward_std": 0.03576334938406944, "rewards/accuracy_reward/mean": 0.017578125, "rewards/accuracy_reward/std": 0.13154059648513794, "step": 760 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.732421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2048.0, "completions/mean_length": 1737.080078125, "completions/mean_terminated_length": 886.0219116210938, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.2597934624903986, "frac_reward_zero_std": 0.78125, "grad_norm": 42.327796975300664, "learning_rate": 9.321321345040391e-07, "loss": 0.002, "num_tokens": 735731308.0, "reward": 0.056640625, "reward_std": 0.08175812661647797, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 761 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.76171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1900.0, "completions/mean_length": 1767.5703125, "completions/mean_terminated_length": 871.1146850585938, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.26013484680378934, "frac_reward_zero_std": 0.875, "grad_norm": 2.326397588053107, "learning_rate": 9.31848745059645e-07, "loss": 0.0057, "num_tokens": 736709936.0, "reward": 0.017578125, "reward_std": 0.04219770431518555, "rewards/accuracy_reward/mean": 0.017578125, "rewards/accuracy_reward/std": 0.13154059648513794, "step": 762 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.771484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2033.0, "completions/mean_length": 1763.228515625, "completions/mean_terminated_length": 801.820556640625, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.26047623111718016, "frac_reward_zero_std": 0.8125, "grad_norm": 3.4820721632697342, "learning_rate": 9.315648136507987e-07, "loss": 0.0114, "num_tokens": 737693173.0, "reward": 0.037109375, "reward_std": 0.058760736137628555, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 763 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1985.0, "completions/mean_length": 1755.501953125, "completions/mean_terminated_length": 800.0083618164062, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.26081761543057097, "frac_reward_zero_std": 0.84375, "grad_norm": 2.966260725492935, "learning_rate": 9.312803406804882e-07, "loss": 0.0101, "num_tokens": 738663350.0, "reward": 0.0234375, "reward_std": 0.05259781330823898, "rewards/accuracy_reward/mean": 0.0234375, "rewards/accuracy_reward/std": 0.15143637359142303, "step": 764 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.740234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1962.0, "completions/mean_length": 1769.802734375, "completions/mean_terminated_length": 977.045166015625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.2611589997439618, "frac_reward_zero_std": 0.8125, "grad_norm": 1.0559549853916272, "learning_rate": 9.309953265524714e-07, "loss": 0.0137, "num_tokens": 739641089.0, "reward": 0.033203125, "reward_std": 0.06570751965045929, "rewards/accuracy_reward/mean": 0.033203125, "rewards/accuracy_reward/std": 0.17934183776378632, "step": 765 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1989.0, "completions/mean_length": 1789.72265625, "completions/mean_terminated_length": 725.9400024414062, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.26150038405735254, "frac_reward_zero_std": 0.8125, "grad_norm": 11.857766669809386, "learning_rate": 9.307097716712735e-07, "loss": 0.0173, "num_tokens": 740641379.0, "reward": 0.037109375, "reward_std": 0.06508206576108932, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 766 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.779296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2011.0, "completions/mean_length": 1796.47265625, "completions/mean_terminated_length": 908.3363037109375, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.26184176837074336, "frac_reward_zero_std": 0.84375, "grad_norm": 19.91026183073299, "learning_rate": 9.304236764421876e-07, "loss": 0.0056, "num_tokens": 741638261.0, "reward": 0.0625, "reward_std": 0.06657323241233826, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.2422981858253479, "step": 767 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.876953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1868.0, "completions/mean_length": 1890.76953125, "completions/mean_terminated_length": 770.1904907226562, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.26218315268413417, "frac_reward_zero_std": 0.90625, "grad_norm": 0.36754778467259164, "learning_rate": 9.301370412712733e-07, "loss": 0.014, "num_tokens": 742681839.0, "reward": 0.01171875, "reward_std": 0.032021719962358475, "rewards/accuracy_reward/mean": 0.012096773833036423, "rewards/accuracy_reward/std": 0.10942844301462173, "step": 768 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.787109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2027.0, "completions/mean_length": 1794.798828125, "completions/mean_terminated_length": 858.6513671875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.262524536997525, "frac_reward_zero_std": 0.84375, "grad_norm": 0.8231903772645666, "learning_rate": 9.29849866565357e-07, "loss": 0.0018, "num_tokens": 743670616.0, "reward": 0.013671875, "reward_std": 0.04478531330823898, "rewards/accuracy_reward/mean": 0.013671875, "rewards/accuracy_reward/std": 0.1162383034825325, "step": 769 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.748046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1963.0, "completions/mean_length": 1745.361328125, "completions/mean_terminated_length": 846.8294677734375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.26286592131091574, "frac_reward_zero_std": 0.84375, "grad_norm": 1.719328430753915, "learning_rate": 9.295621527320305e-07, "loss": 0.0205, "num_tokens": 744645841.0, "reward": 0.05078125, "reward_std": 0.06441686302423477, "rewards/accuracy_reward/mean": 0.052419353276491165, "rewards/accuracy_reward/std": 0.22309619188308716, "step": 770 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1963.0, "completions/mean_length": 1834.23046875, "completions/mean_terminated_length": 858.8043823242188, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.26320730562430655, "frac_reward_zero_std": 0.78125, "grad_norm": 13.02017420406366, "learning_rate": 9.292739001796513e-07, "loss": 0.0147, "num_tokens": 745669111.0, "reward": 0.044921875, "reward_std": 0.07960271090269089, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 771 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.791015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1918.0, "completions/mean_length": 1824.91015625, "completions/mean_terminated_length": 980.504638671875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.26354868993769737, "frac_reward_zero_std": 0.9375, "grad_norm": 1.4715250050149322, "learning_rate": 9.289851093173408e-07, "loss": 0.0063, "num_tokens": 746676265.0, "reward": 0.01171875, "reward_std": 0.024649331346154213, "rewards/accuracy_reward/mean": 0.01171875, "rewards/accuracy_reward/std": 0.10772226005792618, "step": 772 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.830078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1932.0, "completions/mean_length": 1866.5625, "completions/mean_terminated_length": 980.2298583984375, "completions/min_length": 7.0, "completions/min_terminated_length": 7.0, "epoch": 0.2638900742510882, "frac_reward_zero_std": 0.90625, "grad_norm": 2.021038073079869, "learning_rate": 9.286957805549849e-07, "loss": 0.0079, "num_tokens": 747706553.0, "reward": 0.009765625, "reward_std": 0.028222277760505676, "rewards/accuracy_reward/mean": 0.009765625, "rewards/accuracy_reward/std": 0.09843364357948303, "step": 773 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2025.0, "completions/mean_length": 1819.5390625, "completions/mean_terminated_length": 878.2799682617188, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.26423145856447894, "frac_reward_zero_std": 0.78125, "grad_norm": 11.118142475390233, "learning_rate": 9.284059143032329e-07, "loss": 0.0153, "num_tokens": 748716125.0, "reward": 0.056640625, "reward_std": 0.0906703919172287, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 774 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.84765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1955.0, "completions/mean_length": 1870.498046875, "completions/mean_terminated_length": 882.8590087890625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.26457284287786975, "frac_reward_zero_std": 0.90625, "grad_norm": 4.084974035197056, "learning_rate": 9.281155109734971e-07, "loss": 0.0132, "num_tokens": 749745164.0, "reward": 0.029296875, "reward_std": 0.03889618441462517, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 775 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.818359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1981.0, "completions/mean_length": 1850.0546875, "completions/mean_terminated_length": 958.236572265625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.26491422719126057, "frac_reward_zero_std": 0.8125, "grad_norm": 3.6615705296670886, "learning_rate": 9.278245709779515e-07, "loss": 0.011, "num_tokens": 750767096.0, "reward": 0.021484375, "reward_std": 0.059305962175130844, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 776 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.837890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2021.0, "completions/mean_length": 1838.583984375, "completions/mean_terminated_length": 756.1807250976562, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.2652556115046514, "frac_reward_zero_std": 0.90625, "grad_norm": 0.4985521336770881, "learning_rate": 9.275330947295326e-07, "loss": 0.0024, "num_tokens": 751787939.0, "reward": 0.0078125, "reward_std": 0.02629890665411949, "rewards/accuracy_reward/mean": 0.0078125, "rewards/accuracy_reward/std": 0.08812850713729858, "step": 777 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.888671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1848.0, "completions/mean_length": 1932.65625, "completions/mean_terminated_length": 1011.9298095703125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.26559699581804214, "frac_reward_zero_std": 0.875, "grad_norm": 2.796997054538657, "learning_rate": 9.272410826419374e-07, "loss": 0.0079, "num_tokens": 752855379.0, "reward": 0.021484375, "reward_std": 0.04478531330823898, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 778 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.837890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2038.0, "completions/mean_length": 1851.685546875, "completions/mean_terminated_length": 836.9999389648438, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.26593838013143295, "frac_reward_zero_std": 0.875, "grad_norm": 23.433475064428848, "learning_rate": 9.269485351296239e-07, "loss": 0.0043, "num_tokens": 753882178.0, "reward": 0.041015625, "reward_std": 0.05492059141397476, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 779 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.833984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1885.0, "completions/mean_length": 1845.201171875, "completions/mean_terminated_length": 826.435302734375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.26627976444482376, "frac_reward_zero_std": 0.75, "grad_norm": 5.196287934551307, "learning_rate": 9.266554526078095e-07, "loss": 0.0167, "num_tokens": 754899689.0, "reward": 0.03125, "reward_std": 0.07795868813991547, "rewards/accuracy_reward/mean": 0.032258063554763794, "rewards/accuracy_reward/std": 0.17686307430267334, "step": 780 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1773.0, "completions/mean_length": 1850.798828125, "completions/mean_terminated_length": 719.4868774414062, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.2666211487582146, "frac_reward_zero_std": 0.875, "grad_norm": 4.389852331718428, "learning_rate": 9.263618354924714e-07, "loss": 0.0121, "num_tokens": 755920866.0, "reward": 0.015625, "reward_std": 0.03839729726314545, "rewards/accuracy_reward/mean": 0.015625, "rewards/accuracy_reward/std": 0.12414088100194931, "step": 781 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.916015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2041.0, "completions/mean_length": 1944.060546875, "completions/mean_terminated_length": 810.3953247070312, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.26696253307160533, "frac_reward_zero_std": 0.875, "grad_norm": 1.9108395569078425, "learning_rate": 9.260676842003453e-07, "loss": 0.0166, "num_tokens": 756991185.0, "reward": 0.0234375, "reward_std": 0.04522542655467987, "rewards/accuracy_reward/mean": 0.024193547666072845, "rewards/accuracy_reward/std": 0.15380479395389557, "step": 782 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.826171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2044.0, "completions/mean_length": 1835.373046875, "completions/mean_terminated_length": 824.7977905273438, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.26730391738499615, "frac_reward_zero_std": 0.90625, "grad_norm": 7.340623069927433, "learning_rate": 9.257729991489252e-07, "loss": -0.0007, "num_tokens": 758005200.0, "reward": 0.0234375, "reward_std": 0.03748524188995361, "rewards/accuracy_reward/mean": 0.0234375, "rewards/accuracy_reward/std": 0.15143637359142303, "step": 783 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.90625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2014.0, "completions/mean_length": 1921.240234375, "completions/mean_terminated_length": 695.8958740234375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.26764530169838696, "frac_reward_zero_std": 0.875, "grad_norm": 3.7644029241124284, "learning_rate": 9.254777807564626e-07, "loss": 0.0022, "num_tokens": 759063275.0, "reward": 0.01171875, "reward_std": 0.03697281330823898, "rewards/accuracy_reward/mean": 0.01171875, "rewards/accuracy_reward/std": 0.10772226005792618, "step": 784 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.87109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2002.0, "completions/mean_length": 1885.595703125, "completions/mean_terminated_length": 788.1364135742188, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.2679866860117778, "frac_reward_zero_std": 0.90625, "grad_norm": 10.522009974403513, "learning_rate": 9.251820294419661e-07, "loss": 0.0121, "num_tokens": 760101820.0, "reward": 0.03515625, "reward_std": 0.04318207502365112, "rewards/accuracy_reward/mean": 0.03629032149910927, "rewards/accuracy_reward/std": 0.1872003972530365, "step": 785 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.77734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1984.0, "completions/mean_length": 1768.8671875, "completions/mean_terminated_length": 794.3508911132812, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.26832807032516853, "frac_reward_zero_std": 0.78125, "grad_norm": 15.921066859733127, "learning_rate": 9.248857456252005e-07, "loss": 0.0223, "num_tokens": 761080776.0, "reward": 0.046875, "reward_std": 0.08334338665008545, "rewards/accuracy_reward/mean": 0.046875, "rewards/accuracy_reward/std": 0.21157780289649963, "step": 786 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1872.0, "completions/mean_length": 1849.865234375, "completions/mean_terminated_length": 840.3214721679688, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.26866945463855935, "frac_reward_zero_std": 0.90625, "grad_norm": 3.27380743481921, "learning_rate": 9.245889297266866e-07, "loss": 0.016, "num_tokens": 762102323.0, "reward": 0.017578125, "reward_std": 0.03411140665411949, "rewards/accuracy_reward/mean": 0.017578125, "rewards/accuracy_reward/std": 0.13154059648513794, "step": 787 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.943359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2005.0, "completions/mean_length": 1967.6796875, "completions/mean_terminated_length": 629.9310302734375, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.26901083895195016, "frac_reward_zero_std": 0.90625, "grad_norm": 5.748645983839906, "learning_rate": 9.242915821677001e-07, "loss": 0.0108, "num_tokens": 763185279.0, "reward": 0.02734375, "reward_std": 0.039282046258449554, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 788 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1646.0, "completions/mean_length": 1885.955078125, "completions/mean_terminated_length": 827.8970336914062, "completions/min_length": 8.0, "completions/min_terminated_length": 8.0, "epoch": 0.269352223265341, "frac_reward_zero_std": 0.875, "grad_norm": 11.697268262704851, "learning_rate": 9.239937033702717e-07, "loss": 0.0047, "num_tokens": 764227176.0, "reward": 0.029296875, "reward_std": 0.043847277760505676, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 789 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.935546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1402.0, "completions/mean_length": 1957.75, "completions/mean_terminated_length": 647.757568359375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.26969360757873173, "frac_reward_zero_std": 0.9375, "grad_norm": 1.0305707473273247, "learning_rate": 9.236952937571856e-07, "loss": 0.0086, "num_tokens": 765304120.0, "reward": 0.013671875, "reward_std": 0.026572702452540398, "rewards/accuracy_reward/mean": 0.013671875, "rewards/accuracy_reward/std": 0.1162383034825325, "step": 790 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.87890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1982.0, "completions/mean_length": 1908.49609375, "completions/mean_terminated_length": 895.9677124023438, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.27003499189212254, "frac_reward_zero_std": 0.875, "grad_norm": 14.307911798958086, "learning_rate": 9.233963537519799e-07, "loss": 0.0114, "num_tokens": 766346502.0, "reward": 0.02734375, "reward_std": 0.04847268760204315, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 791 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.791015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2037.0, "completions/mean_length": 1809.228515625, "completions/mean_terminated_length": 905.46728515625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.27037637620551336, "frac_reward_zero_std": 0.90625, "grad_norm": 2.3232443725288308, "learning_rate": 9.230968837789451e-07, "loss": 0.0047, "num_tokens": 767349083.0, "reward": 0.009765625, "reward_std": 0.029160313308238983, "rewards/accuracy_reward/mean": 0.009765625, "rewards/accuracy_reward/std": 0.09843364357948303, "step": 792 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.84375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1908.0, "completions/mean_length": 1878.4375, "completions/mean_terminated_length": 962.7999877929688, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.27071776051890417, "frac_reward_zero_std": 0.8125, "grad_norm": 7.687784549686607, "learning_rate": 9.227968842631243e-07, "loss": 0.0235, "num_tokens": 768390571.0, "reward": 0.021484375, "reward_std": 0.059305962175130844, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 793 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.822265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2009.0, "completions/mean_length": 1836.49609375, "completions/mean_terminated_length": 858.6154174804688, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.27105914483229493, "frac_reward_zero_std": 0.90625, "grad_norm": 5.744451020002693, "learning_rate": 9.224963556303116e-07, "loss": 0.0053, "num_tokens": 769420953.0, "reward": 0.029296875, "reward_std": 0.04230976849794388, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 794 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.845703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1931.0, "completions/mean_length": 1871.5390625, "completions/mean_terminated_length": 904.3544311523438, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.27140052914568574, "frac_reward_zero_std": 0.78125, "grad_norm": 5.346002400716161, "learning_rate": 9.221952983070526e-07, "loss": 0.0255, "num_tokens": 770447341.0, "reward": 0.0390625, "reward_std": 0.07961063086986542, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 795 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.85546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1964.0, "completions/mean_length": 1885.09765625, "completions/mean_terminated_length": 920.8919067382812, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.27174191345907656, "frac_reward_zero_std": 0.8125, "grad_norm": 13.155085180387175, "learning_rate": 9.218937127206432e-07, "loss": 0.0058, "num_tokens": 771489167.0, "reward": 0.021484375, "reward_std": 0.06024399772286415, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 796 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.787109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1843.0, "completions/mean_length": 1764.76953125, "completions/mean_terminated_length": 717.5963134765625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.27208329777246737, "frac_reward_zero_std": 0.78125, "grad_norm": 8.939817684934455, "learning_rate": 9.215915992991289e-07, "loss": 0.0225, "num_tokens": 772471321.0, "reward": 0.076171875, "reward_std": 0.08642987906932831, "rewards/accuracy_reward/mean": 0.076171875, "rewards/accuracy_reward/std": 0.26553234457969666, "step": 797 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2013.0, "completions/mean_length": 1884.89453125, "completions/mean_terminated_length": 949.7894897460938, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.2724246820858581, "frac_reward_zero_std": 0.84375, "grad_norm": 12.27218311020411, "learning_rate": 9.212889584713044e-07, "loss": 0.0014, "num_tokens": 773513667.0, "reward": 0.044921875, "reward_std": 0.07057978957891464, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 798 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.927734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2039.0, "completions/mean_length": 1965.400390625, "completions/mean_terminated_length": 905.0, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.27276606639924894, "frac_reward_zero_std": 0.875, "grad_norm": 2.6174194742930634, "learning_rate": 9.20985790666713e-07, "loss": 0.0077, "num_tokens": 774590736.0, "reward": 0.0234375, "reward_std": 0.044233135879039764, "rewards/accuracy_reward/mean": 0.0234375, "rewards/accuracy_reward/std": 0.15143637359142303, "step": 799 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1932.0, "completions/mean_length": 1862.845703125, "completions/mean_terminated_length": 800.6447143554688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.27310745071263975, "frac_reward_zero_std": 0.9375, "grad_norm": 0.06604128369871849, "learning_rate": 9.20682096315646e-07, "loss": 0.007, "num_tokens": 775620417.0, "reward": 0.00390625, "reward_std": 0.015625, "rewards/accuracy_reward/mean": 0.004032257944345474, "rewards/accuracy_reward/std": 0.06343588978052139, "step": 800 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.77734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1951.0, "completions/mean_length": 1792.537109375, "completions/mean_terminated_length": 900.6578979492188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.27344883502603057, "frac_reward_zero_std": 0.875, "grad_norm": 5.199969987735104, "learning_rate": 9.20377875849142e-07, "loss": 0.004, "num_tokens": 776615252.0, "reward": 0.03515625, "reward_std": 0.04175759106874466, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 801 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.85546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1979.0, "completions/mean_length": 1880.59765625, "completions/mean_terminated_length": 889.7567749023438, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.2737902193394214, "frac_reward_zero_std": 0.9375, "grad_norm": 1.2533829936741272, "learning_rate": 9.200731296989862e-07, "loss": 0.0003, "num_tokens": 777654214.0, "reward": 0.0078125, "reward_std": 0.021347813308238983, "rewards/accuracy_reward/mean": 0.0078125, "rewards/accuracy_reward/std": 0.08812850713729858, "step": 802 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.841796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1952.0, "completions/mean_length": 1856.60546875, "completions/mean_terminated_length": 838.197509765625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.27413160365281214, "frac_reward_zero_std": 0.84375, "grad_norm": 6.031500405055657, "learning_rate": 9.1976785829771e-07, "loss": 0.0173, "num_tokens": 778679308.0, "reward": 0.0390625, "reward_std": 0.051493462175130844, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 803 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.837890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1772.0, "completions/mean_length": 1837.734375, "completions/mean_terminated_length": 750.9397583007812, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.27447298796620295, "frac_reward_zero_std": 0.875, "grad_norm": 8.547172002821858, "learning_rate": 9.194620620785905e-07, "loss": 0.0149, "num_tokens": 779692308.0, "reward": 0.025390625, "reward_std": 0.04230976849794388, "rewards/accuracy_reward/mean": 0.025390625, "rewards/accuracy_reward/std": 0.15746226906776428, "step": 804 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1632.0, "completions/mean_length": 1807.783203125, "completions/mean_terminated_length": 909.1944580078125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.27481437227959377, "frac_reward_zero_std": 0.75, "grad_norm": 7.357585515972072, "learning_rate": 9.191557414756495e-07, "loss": 0.0222, "num_tokens": 780690357.0, "reward": 0.06640625, "reward_std": 0.09353180229663849, "rewards/accuracy_reward/mean": 0.06640625, "rewards/accuracy_reward/std": 0.2492343932390213, "step": 805 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.85546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1968.0, "completions/mean_length": 1880.994140625, "completions/mean_terminated_length": 892.5, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.2751557565929846, "frac_reward_zero_std": 0.90625, "grad_norm": 34.122581171970694, "learning_rate": 9.188488969236531e-07, "loss": -0.0032, "num_tokens": 781736514.0, "reward": 0.03515625, "reward_std": 0.040593504905700684, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 806 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.81640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1841.494140625, "completions/mean_terminated_length": 923.7340087890625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.27549714090637534, "frac_reward_zero_std": 0.75, "grad_norm": 25.69693549648914, "learning_rate": 9.185415288581105e-07, "loss": -0.0018, "num_tokens": 782762143.0, "reward": 0.03125, "reward_std": 0.08257714658975601, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 807 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.822265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1970.0, "completions/mean_length": 1846.7421875, "completions/mean_terminated_length": 915.6483764648438, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.27583852521976615, "frac_reward_zero_std": 0.75, "grad_norm": 8.414492313316417, "learning_rate": 9.182336377152753e-07, "loss": 0.0004, "num_tokens": 783782891.0, "reward": 0.0703125, "reward_std": 0.0937047079205513, "rewards/accuracy_reward/mean": 0.0703125, "rewards/accuracy_reward/std": 0.25592297315597534, "step": 808 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.732421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2036.0, "completions/mean_length": 1761.166015625, "completions/mean_terminated_length": 976.0364990234375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.27617990953315696, "frac_reward_zero_std": 0.78125, "grad_norm": 4.193371531586765, "learning_rate": 9.179252239321419e-07, "loss": 0.0082, "num_tokens": 784769984.0, "reward": 0.09375, "reward_std": 0.09039659798145294, "rewards/accuracy_reward/mean": 0.09375, "rewards/accuracy_reward/std": 0.29176566004753113, "step": 809 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.80078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2008.0, "completions/mean_length": 1808.0625, "completions/mean_terminated_length": 843.6078491210938, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.2765212938465478, "frac_reward_zero_std": 0.90625, "grad_norm": 6.400832735733385, "learning_rate": 9.176162879464477e-07, "loss": 0.0034, "num_tokens": 785769088.0, "reward": 0.03125, "reward_std": 0.03996079042553902, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 810 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.86328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2045.0, "completions/mean_length": 1880.482421875, "completions/mean_terminated_length": 822.7285766601562, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.27686267815993854, "frac_reward_zero_std": 0.875, "grad_norm": 1.1016606703853062, "learning_rate": 9.173068301966707e-07, "loss": 0.0115, "num_tokens": 786808935.0, "reward": 0.0234375, "reward_std": 0.03944835811853409, "rewards/accuracy_reward/mean": 0.024193547666072845, "rewards/accuracy_reward/std": 0.15380479395389557, "step": 811 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.720703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1974.0, "completions/mean_length": 1691.0078125, "completions/mean_terminated_length": 769.8181762695312, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.27720406247332935, "frac_reward_zero_std": 0.6875, "grad_norm": 5.374550915557742, "learning_rate": 9.169968511220296e-07, "loss": 0.0093, "num_tokens": 787754555.0, "reward": 0.0625, "reward_std": 0.10909565538167953, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.2422981858253479, "step": 812 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.767578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1935.0, "completions/mean_length": 1773.072265625, "completions/mean_terminated_length": 865.11767578125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.27754544678672016, "frac_reward_zero_std": 0.9375, "grad_norm": 6.333591138045338, "learning_rate": 9.166863511624828e-07, "loss": 0.0038, "num_tokens": 788739920.0, "reward": 0.01953125, "reward_std": 0.02681133709847927, "rewards/accuracy_reward/mean": 0.01953125, "rewards/accuracy_reward/std": 0.1385180652141571, "step": 813 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1992.0, "completions/mean_length": 1762.4609375, "completions/mean_terminated_length": 829.7000732421875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.277886831100111, "frac_reward_zero_std": 0.75, "grad_norm": 1.7085857978749721, "learning_rate": 9.163753307587285e-07, "loss": 0.0251, "num_tokens": 789715884.0, "reward": 0.029296875, "reward_std": 0.07729348540306091, "rewards/accuracy_reward/mean": 0.030241934582591057, "rewards/accuracy_reward/std": 0.1714252382516861, "step": 814 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.771484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2029.0, "completions/mean_length": 1802.287109375, "completions/mean_terminated_length": 972.74365234375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.27822821541350173, "frac_reward_zero_std": 0.8125, "grad_norm": 14.597076921488581, "learning_rate": 9.160637903522031e-07, "loss": 0.0082, "num_tokens": 790720943.0, "reward": 0.044921875, "reward_std": 0.06684703379869461, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 815 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1927.0, "completions/mean_length": 1742.84375, "completions/mean_terminated_length": 932.0, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.27856959972689255, "frac_reward_zero_std": 0.78125, "grad_norm": 4.281868494887167, "learning_rate": 9.157517303850814e-07, "loss": 0.0058, "num_tokens": 791696863.0, "reward": 0.0390625, "reward_std": 0.07289456576108932, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 816 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2045.0, "completions/mean_length": 1767.712890625, "completions/mean_terminated_length": 852.1083984375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.27891098404028336, "frac_reward_zero_std": 0.9375, "grad_norm": 3.919495329929373, "learning_rate": 9.154391513002754e-07, "loss": 0.0052, "num_tokens": 792677932.0, "reward": 0.0078125, "reward_std": 0.021347813308238983, "rewards/accuracy_reward/mean": 0.0078125, "rewards/accuracy_reward/std": 0.08812850713729858, "step": 817 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.830078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1949.0, "completions/mean_length": 1842.521484375, "completions/mean_terminated_length": 838.7471313476562, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.2792523683536742, "frac_reward_zero_std": 0.75, "grad_norm": 6.57673311161995, "learning_rate": 9.151260535414336e-07, "loss": 0.015, "num_tokens": 793694375.0, "reward": 0.037109375, "reward_std": 0.08133251965045929, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 818 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.806640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2034.0, "completions/mean_length": 1821.9453125, "completions/mean_terminated_length": 878.9091186523438, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.27959375266706493, "frac_reward_zero_std": 0.75, "grad_norm": 10.636962021839514, "learning_rate": 9.148124375529414e-07, "loss": 0.0219, "num_tokens": 794708299.0, "reward": 0.046875, "reward_std": 0.09228022396564484, "rewards/accuracy_reward/mean": 0.046875, "rewards/accuracy_reward/std": 0.21157780289649963, "step": 819 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.771484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2012.0, "completions/mean_length": 1774.3828125, "completions/mean_terminated_length": 850.6325073242188, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.27993513698045575, "frac_reward_zero_std": 0.8125, "grad_norm": 4.16909340269041, "learning_rate": 9.144983037799192e-07, "loss": 0.0134, "num_tokens": 795689279.0, "reward": 0.02734375, "reward_std": 0.0635918527841568, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 820 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.724609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1761.0, "completions/mean_length": 1698.392578125, "completions/mean_terminated_length": 778.5035400390625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.28027652129384656, "frac_reward_zero_std": 0.84375, "grad_norm": 6.886377386385229, "learning_rate": 9.141836526682226e-07, "loss": 0.0175, "num_tokens": 796633400.0, "reward": 0.05859375, "reward_std": 0.055899329483509064, "rewards/accuracy_reward/mean": 0.05859375, "rewards/accuracy_reward/std": 0.23509246110916138, "step": 821 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.767578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1901.0, "completions/mean_length": 1802.443359375, "completions/mean_terminated_length": 991.4874267578125, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.2806179056072374, "frac_reward_zero_std": 0.84375, "grad_norm": 10.070686773266614, "learning_rate": 9.138684846644408e-07, "loss": 0.0091, "num_tokens": 797637691.0, "reward": 0.03125, "reward_std": 0.05259781330823898, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 822 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.740234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2045.0, "completions/mean_length": 1741.353515625, "completions/mean_terminated_length": 867.5263671875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.28095928992062813, "frac_reward_zero_std": 0.78125, "grad_norm": 10.327527105571596, "learning_rate": 9.135528002158977e-07, "loss": -0.0003, "num_tokens": 798608704.0, "reward": 0.083984375, "reward_std": 0.0775209367275238, "rewards/accuracy_reward/mean": 0.083984375, "rewards/accuracy_reward/std": 0.2776356339454651, "step": 823 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.740234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2024.0, "completions/mean_length": 1723.576171875, "completions/mean_terminated_length": 799.0902709960938, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.28130067423401894, "frac_reward_zero_std": 0.875, "grad_norm": 1.5841398569998713, "learning_rate": 9.132365997706493e-07, "loss": 0.0238, "num_tokens": 799573847.0, "reward": 0.03515625, "reward_std": 0.056218504905700684, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 824 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.830078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1994.0, "completions/mean_length": 1835.443359375, "completions/mean_terminated_length": 797.0919799804688, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.28164205854740976, "frac_reward_zero_std": 0.90625, "grad_norm": 1.6252711039686722, "learning_rate": 9.129198837774846e-07, "loss": 0.0149, "num_tokens": 800590362.0, "reward": 0.0234375, "reward_std": 0.03125, "rewards/accuracy_reward/mean": 0.0234375, "rewards/accuracy_reward/std": 0.15143637359142303, "step": 825 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.763671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2021.0, "completions/mean_length": 1760.51953125, "completions/mean_terminated_length": 831.5537109375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.28198344286080057, "frac_reward_zero_std": 0.90625, "grad_norm": 2.255931144584967, "learning_rate": 9.126026526859236e-07, "loss": 0.0088, "num_tokens": 801566676.0, "reward": 0.01953125, "reward_std": 0.03674772381782532, "rewards/accuracy_reward/mean": 0.01953125, "rewards/accuracy_reward/std": 0.1385180652141571, "step": 826 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.751953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2029.0, "completions/mean_length": 1764.47265625, "completions/mean_terminated_length": 905.3228149414062, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.28232482717419133, "frac_reward_zero_std": 0.8125, "grad_norm": 6.973680788850989, "learning_rate": 9.122849069462181e-07, "loss": 0.0094, "num_tokens": 802562966.0, "reward": 0.12890625, "reward_std": 0.06794347614049911, "rewards/accuracy_reward/mean": 0.12890625, "rewards/accuracy_reward/std": 0.33542385697364807, "step": 827 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.791015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1983.0, "completions/mean_length": 1803.73046875, "completions/mean_terminated_length": 879.1588745117188, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.28266621148758214, "frac_reward_zero_std": 0.90625, "grad_norm": 0.7959328963289285, "learning_rate": 9.119666470093501e-07, "loss": 0.0198, "num_tokens": 803566636.0, "reward": 0.03515625, "reward_std": 0.035868462175130844, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 828 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.771484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2043.0, "completions/mean_length": 1765.677734375, "completions/mean_terminated_length": 812.5385131835938, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.28300759580097296, "frac_reward_zero_std": 0.8125, "grad_norm": 4.502840958414974, "learning_rate": 9.116478733270312e-07, "loss": 0.0072, "num_tokens": 804550391.0, "reward": 0.01953125, "reward_std": 0.05402229726314545, "rewards/accuracy_reward/mean": 0.01953125, "rewards/accuracy_reward/std": 0.1385180652141571, "step": 829 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.86328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2019.0, "completions/mean_length": 1896.322265625, "completions/mean_terminated_length": 938.585693359375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.28334898011436377, "frac_reward_zero_std": 0.84375, "grad_norm": 12.189984021375238, "learning_rate": 9.113285863517024e-07, "loss": 0.0199, "num_tokens": 805603068.0, "reward": 0.03125, "reward_std": 0.05259781330823898, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 830 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.876953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1976.0, "completions/mean_length": 1907.384765625, "completions/mean_terminated_length": 905.2222900390625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.2836903644277545, "frac_reward_zero_std": 0.9375, "grad_norm": 3.395821068244509, "learning_rate": 9.110087865365333e-07, "loss": 0.0049, "num_tokens": 806654353.0, "reward": 0.0078125, "reward_std": 0.020409777760505676, "rewards/accuracy_reward/mean": 0.0078125, "rewards/accuracy_reward/std": 0.08812850713729858, "step": 831 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.85546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1791.0, "completions/mean_length": 1863.318359375, "completions/mean_terminated_length": 770.9459838867188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.28403174874114534, "frac_reward_zero_std": 0.75, "grad_norm": 20.836372449782818, "learning_rate": 9.10688474335421e-07, "loss": 0.0026, "num_tokens": 807689348.0, "reward": 0.060546875, "reward_std": 0.08780106902122498, "rewards/accuracy_reward/mean": 0.06458333134651184, "rewards/accuracy_reward/std": 0.24604564905166626, "step": 832 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2037.0, "completions/mean_length": 1843.39453125, "completions/mean_terminated_length": 957.3021240234375, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.28437313305453615, "frac_reward_zero_std": 0.75, "grad_norm": 17.01190114557995, "learning_rate": 9.103676502029901e-07, "loss": -0.0058, "num_tokens": 808725454.0, "reward": 0.0625, "reward_std": 0.10064490139484406, "rewards/accuracy_reward/mean": 0.06451612710952759, "rewards/accuracy_reward/std": 0.2459181249141693, "step": 833 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.837890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1955.0, "completions/mean_length": 1839.3046875, "completions/mean_terminated_length": 760.62646484375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.28471451736792697, "frac_reward_zero_std": 0.84375, "grad_norm": 16.371139320984373, "learning_rate": 9.100463145945921e-07, "loss": 0.0225, "num_tokens": 809750762.0, "reward": 0.0390625, "reward_std": 0.062054343521595, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 834 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.865234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1839.0, "completions/mean_length": 1884.580078125, "completions/mean_terminated_length": 835.3768310546875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.2850559016813177, "frac_reward_zero_std": 0.75, "grad_norm": 4.270470032206402, "learning_rate": 9.097244679663037e-07, "loss": 0.0174, "num_tokens": 810792291.0, "reward": 0.0390625, "reward_std": 0.08544550836086273, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 835 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.865234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2041.0, "completions/mean_length": 1904.75, "completions/mean_terminated_length": 985.0435180664062, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.28539728599470854, "frac_reward_zero_std": 0.78125, "grad_norm": 4.193138502196356, "learning_rate": 9.094021107749277e-07, "loss": 0.0265, "num_tokens": 811846467.0, "reward": 0.03125, "reward_std": 0.07426576316356659, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 836 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.880859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1912.0, "completions/mean_length": 1888.322265625, "completions/mean_terminated_length": 707.7540283203125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.28573867030809935, "frac_reward_zero_std": 0.8125, "grad_norm": 8.403438460241839, "learning_rate": 9.090792434779911e-07, "loss": 0.0234, "num_tokens": 812888600.0, "reward": 0.04296875, "reward_std": 0.07443207502365112, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 837 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.822265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1823.0, "completions/mean_length": 1829.146484375, "completions/mean_terminated_length": 816.6483764648438, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.28608005462149017, "frac_reward_zero_std": 0.84375, "grad_norm": 11.672956930438959, "learning_rate": 9.087558665337447e-07, "loss": 0.0038, "num_tokens": 813903699.0, "reward": 0.01171875, "reward_std": 0.04192390665411949, "rewards/accuracy_reward/mean": 0.01171875, "rewards/accuracy_reward/std": 0.10772226005792618, "step": 838 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.822265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1997.0, "completions/mean_length": 1836.85546875, "completions/mean_terminated_length": 860.02197265625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.2864214389348809, "frac_reward_zero_std": 0.8125, "grad_norm": 3.2238850401099515, "learning_rate": 9.084319804011631e-07, "loss": 0.0234, "num_tokens": 814915257.0, "reward": 0.05859375, "reward_std": 0.0756574496626854, "rewards/accuracy_reward/mean": 0.05859375, "rewards/accuracy_reward/std": 0.23509246110916138, "step": 839 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.87109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2043.0, "completions/mean_length": 1878.810546875, "completions/mean_terminated_length": 735.5, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.28676282324827174, "frac_reward_zero_std": 0.9375, "grad_norm": 4.127142651197428, "learning_rate": 9.081075855399434e-07, "loss": 0.0043, "num_tokens": 815950984.0, "reward": 0.005859375, "reward_std": 0.01848640665411949, "rewards/accuracy_reward/mean": 0.005859375, "rewards/accuracy_reward/std": 0.07639661431312561, "step": 840 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1771.0, "completions/mean_length": 1939.169921875, "completions/mean_terminated_length": 976.4423217773438, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.28710420756166255, "frac_reward_zero_std": 0.875, "grad_norm": 2.683663873918988, "learning_rate": 9.077826824105049e-07, "loss": 0.0153, "num_tokens": 817018511.0, "reward": 0.0078125, "reward_std": 0.03125, "rewards/accuracy_reward/mean": 0.0078125, "rewards/accuracy_reward/std": 0.08812850713729858, "step": 841 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.9140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1683.0, "completions/mean_length": 1942.5234375, "completions/mean_terminated_length": 820.6364135742188, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.28744559187505336, "frac_reward_zero_std": 0.8125, "grad_norm": 1.423387792507175, "learning_rate": 9.074572714739881e-07, "loss": 0.023, "num_tokens": 818080715.0, "reward": 0.01953125, "reward_std": 0.05644455552101135, "rewards/accuracy_reward/mean": 0.01953125, "rewards/accuracy_reward/std": 0.1385180652141571, "step": 842 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.83984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2025.0, "completions/mean_length": 1834.3828125, "completions/mean_terminated_length": 714.195068359375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.2877869761884441, "frac_reward_zero_std": 0.875, "grad_norm": 3.914633061898735, "learning_rate": 9.071313531922541e-07, "loss": 0.0119, "num_tokens": 819100495.0, "reward": 0.03125, "reward_std": 0.04847268760204315, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 843 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.861328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2026.0, "completions/mean_length": 1890.33984375, "completions/mean_terminated_length": 911.0704345703125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.28812836050183493, "frac_reward_zero_std": 0.78125, "grad_norm": 18.375183140345467, "learning_rate": 9.068049280278846e-07, "loss": -0.0024, "num_tokens": 820145677.0, "reward": 0.064453125, "reward_std": 0.09178133308887482, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 844 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.853515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1857.0, "completions/mean_length": 1858.984375, "completions/mean_terminated_length": 757.6533813476562, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.28846974481522575, "frac_reward_zero_std": 0.90625, "grad_norm": 4.691095628905756, "learning_rate": 9.064779964441802e-07, "loss": 0.0059, "num_tokens": 821172613.0, "reward": 0.021484375, "reward_std": 0.036420635879039764, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 845 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.79296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1943.0, "completions/mean_length": 1781.109375, "completions/mean_terminated_length": 758.867919921875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.28881112912861656, "frac_reward_zero_std": 0.78125, "grad_norm": 6.227706853654657, "learning_rate": 9.061505589051606e-07, "loss": 0.0261, "num_tokens": 822157597.0, "reward": 0.06640625, "reward_std": 0.0843944400548935, "rewards/accuracy_reward/mean": 0.06640625, "rewards/accuracy_reward/std": 0.2492343932390213, "step": 846 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.873046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1661.0, "completions/mean_length": 1893.7109375, "completions/mean_terminated_length": 832.6769409179688, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.2891525134420073, "frac_reward_zero_std": 0.78125, "grad_norm": 13.660740416830857, "learning_rate": 9.058226158755634e-07, "loss": 0.0207, "num_tokens": 823202553.0, "reward": 0.0234375, "reward_std": 0.06805649399757385, "rewards/accuracy_reward/mean": 0.0234375, "rewards/accuracy_reward/std": 0.15143637359142303, "step": 847 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.83984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1983.0, "completions/mean_length": 1854.662109375, "completions/mean_terminated_length": 840.8170166015625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.28949389775539813, "frac_reward_zero_std": 0.8125, "grad_norm": 12.978697093397896, "learning_rate": 9.05494167820844e-07, "loss": 0.0011, "num_tokens": 824236076.0, "reward": 0.041015625, "reward_std": 0.069866843521595, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 848 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.845703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1964.0, "completions/mean_length": 1848.302734375, "completions/mean_terminated_length": 753.759521484375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.28983528206878895, "frac_reward_zero_std": 0.84375, "grad_norm": 2.294131623082693, "learning_rate": 9.05165215207174e-07, "loss": 0.009, "num_tokens": 825258343.0, "reward": 0.0234375, "reward_std": 0.053855981677770615, "rewards/accuracy_reward/mean": 0.0234375, "rewards/accuracy_reward/std": 0.15143637359142303, "step": 849 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.849609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1835.0, "completions/mean_length": 1860.3359375, "completions/mean_terminated_length": 800.1558227539062, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.29017666638217976, "frac_reward_zero_std": 0.875, "grad_norm": 14.250455706685674, "learning_rate": 9.048357585014417e-07, "loss": 0.0123, "num_tokens": 826290099.0, "reward": 0.048828125, "reward_std": 0.06075643002986908, "rewards/accuracy_reward/mean": 0.05040322616696358, "rewards/accuracy_reward/std": 0.21899642050266266, "step": 850 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.837890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1983.0, "completions/mean_length": 1866.08203125, "completions/mean_terminated_length": 925.8071899414062, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.2905180506955705, "frac_reward_zero_std": 0.875, "grad_norm": 4.8172943673488975, "learning_rate": 9.045057981712504e-07, "loss": 0.0015, "num_tokens": 827324605.0, "reward": 0.017578125, "reward_std": 0.0390625, "rewards/accuracy_reward/mean": 0.017578125, "rewards/accuracy_reward/std": 0.13154059648513794, "step": 851 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1721.0, "completions/mean_length": 1782.380859375, "completions/mean_terminated_length": 740.3365478515625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.29085943500896133, "frac_reward_zero_std": 0.78125, "grad_norm": 12.221566492598393, "learning_rate": 9.041753346849187e-07, "loss": 0.0073, "num_tokens": 828319232.0, "reward": 0.05078125, "reward_std": 0.08874702453613281, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 852 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2045.0, "completions/mean_length": 1754.94140625, "completions/mean_terminated_length": 797.61669921875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.29120081932235214, "frac_reward_zero_std": 0.8125, "grad_norm": 18.8383758396224, "learning_rate": 9.038443685114791e-07, "loss": 0.0212, "num_tokens": 829292546.0, "reward": 0.0625, "reward_std": 0.07708083093166351, "rewards/accuracy_reward/mean": 0.06451612710952759, "rewards/accuracy_reward/std": 0.2459181249141693, "step": 853 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.84765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1960.0, "completions/mean_length": 1877.115234375, "completions/mean_terminated_length": 926.2948608398438, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.29154220363574296, "frac_reward_zero_std": 0.875, "grad_norm": 4.906759911723979, "learning_rate": 9.035129001206771e-07, "loss": -0.0017, "num_tokens": 830320605.0, "reward": 0.021484375, "reward_std": 0.04604348540306091, "rewards/accuracy_reward/mean": 0.02217741869390011, "rewards/accuracy_reward/std": 0.14740893244743347, "step": 854 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1935.0, "completions/mean_length": 1752.626953125, "completions/mean_terminated_length": 828.3951416015625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.2918835879491337, "frac_reward_zero_std": 0.8125, "grad_norm": 17.75494572292462, "learning_rate": 9.03180929982972e-07, "loss": -0.0093, "num_tokens": 831292446.0, "reward": 0.033203125, "reward_std": 0.06661957502365112, "rewards/accuracy_reward/mean": 0.033203125, "rewards/accuracy_reward/std": 0.17934183776378632, "step": 855 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.767578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1990.0, "completions/mean_length": 1738.087890625, "completions/mean_terminated_length": 714.5966796875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.29222497226252453, "frac_reward_zero_std": 0.84375, "grad_norm": 7.0996767404272845, "learning_rate": 9.028484585695345e-07, "loss": 0.0104, "num_tokens": 832261019.0, "reward": 0.044921875, "reward_std": 0.048086829483509064, "rewards/accuracy_reward/mean": 0.0463709682226181, "rewards/accuracy_reward/std": 0.21049949526786804, "step": 856 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.791015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1903.0, "completions/mean_length": 1798.583984375, "completions/mean_terminated_length": 854.8971557617188, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.29256635657591534, "frac_reward_zero_std": 0.78125, "grad_norm": 12.859577799381787, "learning_rate": 9.025154863522467e-07, "loss": 0.0035, "num_tokens": 833266534.0, "reward": 0.05859375, "reward_std": 0.07944335043430328, "rewards/accuracy_reward/mean": 0.05859375, "rewards/accuracy_reward/std": 0.23509246110916138, "step": 857 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.748046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1884.0, "completions/mean_length": 1762.826171875, "completions/mean_terminated_length": 916.1472778320312, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.29290774088930616, "frac_reward_zero_std": 0.90625, "grad_norm": 6.829967668780411, "learning_rate": 9.021820138037022e-07, "loss": 0.0053, "num_tokens": 834241837.0, "reward": 0.0078125, "reward_std": 0.02629890665411949, "rewards/accuracy_reward/mean": 0.008064515888690948, "rewards/accuracy_reward/std": 0.0895301103591919, "step": 858 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.755859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1946.0, "completions/mean_length": 1745.716796875, "completions/mean_terminated_length": 809.8480224609375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.2932491252026969, "frac_reward_zero_std": 0.875, "grad_norm": 14.248401421965156, "learning_rate": 9.01848041397204e-07, "loss": 0.01, "num_tokens": 835214332.0, "reward": 0.041015625, "reward_std": 0.04125870764255524, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 859 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1846.0, "completions/mean_length": 1802.109375, "completions/mean_terminated_length": 837.4615478515625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.2935905095160877, "frac_reward_zero_std": 0.8125, "grad_norm": 8.343051079539533, "learning_rate": 9.015135696067649e-07, "loss": 0.0016, "num_tokens": 836209524.0, "reward": 0.060546875, "reward_std": 0.06849660724401474, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.24230584502220154, "step": 860 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.806640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2027.0, "completions/mean_length": 1837.31640625, "completions/mean_terminated_length": 958.404052734375, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.29393189382947854, "frac_reward_zero_std": 0.75, "grad_norm": 12.252634290375203, "learning_rate": 9.011785989071066e-07, "loss": 0.0281, "num_tokens": 837220630.0, "reward": 0.060546875, "reward_std": 0.09016455709934235, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 861 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.84765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1985.0, "completions/mean_length": 1874.43359375, "completions/mean_terminated_length": 908.6923217773438, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.29427327814286935, "frac_reward_zero_std": 0.875, "grad_norm": 11.947210803084694, "learning_rate": 9.008431297736585e-07, "loss": 0.0112, "num_tokens": 838256372.0, "reward": 0.04296875, "reward_std": 0.040274329483509064, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 862 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.798828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1985.0, "completions/mean_length": 1803.681640625, "completions/mean_terminated_length": 833.5242919921875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.2946146624562601, "frac_reward_zero_std": 0.875, "grad_norm": 11.315930331018635, "learning_rate": 9.005071626825577e-07, "loss": 0.0137, "num_tokens": 839252321.0, "reward": 0.017578125, "reward_std": 0.04125870764255524, "rewards/accuracy_reward/mean": 0.017578125, "rewards/accuracy_reward/std": 0.13154059648513794, "step": 863 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2035.0, "completions/mean_length": 1866.99609375, "completions/mean_terminated_length": 760.8611450195312, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.2949560467696509, "frac_reward_zero_std": 0.875, "grad_norm": 3.6635401294550776, "learning_rate": 9.001706981106482e-07, "loss": 0.0048, "num_tokens": 840288111.0, "reward": 0.015625, "reward_std": 0.040274329483509064, "rewards/accuracy_reward/mean": 0.015625, "rewards/accuracy_reward/std": 0.12414088100194931, "step": 864 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.767578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.0, "completions/mean_length": 1771.62109375, "completions/mean_terminated_length": 858.8740234375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.29529743108304174, "frac_reward_zero_std": 0.8125, "grad_norm": 21.12347243598158, "learning_rate": 8.998337365354798e-07, "loss": 0.0147, "num_tokens": 841273037.0, "reward": 0.02734375, "reward_std": 0.05806133523583412, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 865 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.75, "completions/max_length": 2048.0, "completions/max_terminated_length": 1983.0, "completions/mean_length": 1755.48046875, "completions/mean_terminated_length": 877.921875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.29563881539643255, "frac_reward_zero_std": 0.78125, "grad_norm": 5.274812844600458, "learning_rate": 8.994962784353079e-07, "loss": 0.0226, "num_tokens": 842246931.0, "reward": 0.04296875, "reward_std": 0.08157937228679657, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 866 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.787109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1955.0, "completions/mean_length": 1800.8828125, "completions/mean_terminated_length": 887.2293090820312, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.2959801997098233, "frac_reward_zero_std": 0.90625, "grad_norm": 2.2190727334438605, "learning_rate": 8.991583242890924e-07, "loss": 0.0025, "num_tokens": 843253927.0, "reward": 0.017578125, "reward_std": 0.03536957502365112, "rewards/accuracy_reward/mean": 0.017578125, "rewards/accuracy_reward/std": 0.13154059648513794, "step": 867 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.79296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1988.0, "completions/mean_length": 1804.8125, "completions/mean_terminated_length": 873.3585205078125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.2963215840232141, "frac_reward_zero_std": 0.90625, "grad_norm": 6.815472197270014, "learning_rate": 8.988198745764976e-07, "loss": 0.0065, "num_tokens": 844263143.0, "reward": 0.021484375, "reward_std": 0.036420635879039764, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 868 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.837890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1814.0, "completions/mean_length": 1864.06640625, "completions/mean_terminated_length": 913.3734741210938, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.29666296833660494, "frac_reward_zero_std": 0.9375, "grad_norm": 2.3164676650613054, "learning_rate": 8.984809297778908e-07, "loss": 0.0075, "num_tokens": 845294409.0, "reward": 0.03125, "reward_std": 0.015625, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 869 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.841796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2020.0, "completions/mean_length": 1878.345703125, "completions/mean_terminated_length": 975.6173095703125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.29700435264999575, "frac_reward_zero_std": 0.71875, "grad_norm": 10.956686543784919, "learning_rate": 8.981414903743423e-07, "loss": 0.0294, "num_tokens": 846325562.0, "reward": 0.0625, "reward_std": 0.09671888500452042, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.2422981858253479, "step": 870 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1979.0, "completions/mean_length": 1888.41796875, "completions/mean_terminated_length": 972.9210815429688, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.2973457369633865, "frac_reward_zero_std": 0.90625, "grad_norm": 5.498690871532542, "learning_rate": 8.97801556847624e-07, "loss": 0.0097, "num_tokens": 847370768.0, "reward": 0.021484375, "reward_std": 0.03163585811853409, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 871 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2024.0, "completions/mean_length": 1848.6328125, "completions/mean_terminated_length": 938.478271484375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.2976871212767773, "frac_reward_zero_std": 0.875, "grad_norm": 5.730046789814379, "learning_rate": 8.974611296802096e-07, "loss": 0.0024, "num_tokens": 848394580.0, "reward": 0.046875, "reward_std": 0.044233135879039764, "rewards/accuracy_reward/mean": 0.046875, "rewards/accuracy_reward/std": 0.21157780289649963, "step": 872 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.857421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1894.0, "completions/mean_length": 1881.431640625, "completions/mean_terminated_length": 879.73974609375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.29802850559016814, "frac_reward_zero_std": 0.875, "grad_norm": 5.7773680058847106, "learning_rate": 8.971202093552731e-07, "loss": 0.021, "num_tokens": 849434705.0, "reward": 0.03515625, "reward_std": 0.05138043686747551, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 873 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2044.0, "completions/mean_length": 1808.109375, "completions/mean_terminated_length": 867.0000610351562, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.29836988990355895, "frac_reward_zero_std": 0.875, "grad_norm": 1.7376704146901976, "learning_rate": 8.967787963566887e-07, "loss": 0.0124, "num_tokens": 850433001.0, "reward": 0.025390625, "reward_std": 0.04620979726314545, "rewards/accuracy_reward/mean": 0.025390625, "rewards/accuracy_reward/std": 0.15746226906776428, "step": 874 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2043.0, "completions/mean_length": 1800.59375, "completions/mean_terminated_length": 875.1111450195312, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.2987112742169497, "frac_reward_zero_std": 0.78125, "grad_norm": 14.896390218321304, "learning_rate": 8.964368911690296e-07, "loss": 0.0097, "num_tokens": 851447129.0, "reward": 0.04296875, "reward_std": 0.07317391037940979, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 875 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.78125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1904.0, "completions/mean_length": 1783.162109375, "completions/mean_terminated_length": 837.3125610351562, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.2990526585303405, "frac_reward_zero_std": 0.78125, "grad_norm": 12.428323802598927, "learning_rate": 8.960944942775675e-07, "loss": 0.027, "num_tokens": 852433180.0, "reward": 0.078125, "reward_std": 0.09060370922088623, "rewards/accuracy_reward/mean": 0.078125, "rewards/accuracy_reward/std": 0.26863065361976624, "step": 876 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.79296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1909.0, "completions/mean_length": 1785.802734375, "completions/mean_terminated_length": 781.5377197265625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.29939404284373133, "frac_reward_zero_std": 0.875, "grad_norm": 5.528397245482416, "learning_rate": 8.957516061682724e-07, "loss": 0.016, "num_tokens": 853427783.0, "reward": 0.029296875, "reward_std": 0.049457065761089325, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 877 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.794921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2029.0, "completions/mean_length": 1812.375, "completions/mean_terminated_length": 899.0476684570312, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.29973542715712215, "frac_reward_zero_std": 0.875, "grad_norm": 10.188880628697577, "learning_rate": 8.954082273278112e-07, "loss": 0.0094, "num_tokens": 854430631.0, "reward": 0.029296875, "reward_std": 0.049457065761089325, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 878 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.744140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2002.0, "completions/mean_length": 1742.98828125, "completions/mean_terminated_length": 856.572509765625, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.3000768114705129, "frac_reward_zero_std": 0.71875, "grad_norm": 11.391181433833179, "learning_rate": 8.950643582435474e-07, "loss": 0.0102, "num_tokens": 855410689.0, "reward": 0.05859375, "reward_std": 0.10986974090337753, "rewards/accuracy_reward/mean": 0.060483869165182114, "rewards/accuracy_reward/std": 0.2386218160390854, "step": 879 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.748046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2034.0, "completions/mean_length": 1748.1953125, "completions/mean_terminated_length": 858.465087890625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.3004181957839037, "frac_reward_zero_std": 0.84375, "grad_norm": 17.016757797584134, "learning_rate": 8.9471999940354e-07, "loss": 0.0125, "num_tokens": 856402005.0, "reward": 0.033203125, "reward_std": 0.06018522381782532, "rewards/accuracy_reward/mean": 0.033203125, "rewards/accuracy_reward/std": 0.17934183776378632, "step": 880 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.76953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2041.0, "completions/mean_length": 1779.482421875, "completions/mean_terminated_length": 882.9067993164062, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.30075958009729453, "frac_reward_zero_std": 0.71875, "grad_norm": 13.073247218031963, "learning_rate": 8.943751512965437e-07, "loss": 0.0415, "num_tokens": 857384876.0, "reward": 0.060546875, "reward_std": 0.10088353604078293, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 881 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.814453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1923.0, "completions/mean_length": 1826.751953125, "completions/mean_terminated_length": 855.5895385742188, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.30110096441068535, "frac_reward_zero_std": 0.75, "grad_norm": 9.278548574437648, "learning_rate": 8.940298144120074e-07, "loss": 0.0236, "num_tokens": 858399885.0, "reward": 0.05859375, "reward_std": 0.10272903740406036, "rewards/accuracy_reward/mean": 0.05859375, "rewards/accuracy_reward/std": 0.23509246110916138, "step": 882 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.810546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1952.0, "completions/mean_length": 1852.19140625, "completions/mean_terminated_length": 1014.4535522460938, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.3014423487240761, "frac_reward_zero_std": 0.65625, "grad_norm": 3.58069200787596, "learning_rate": 8.936839892400732e-07, "loss": 0.0297, "num_tokens": 859427023.0, "reward": 0.0546875, "reward_std": 0.11459337919950485, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 883 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2042.0, "completions/mean_length": 1802.20703125, "completions/mean_terminated_length": 882.75927734375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3017837330374669, "frac_reward_zero_std": 0.84375, "grad_norm": 6.840636193515991, "learning_rate": 8.93337676271577e-07, "loss": -0.006, "num_tokens": 860428121.0, "reward": 0.041015625, "reward_std": 0.06750431656837463, "rewards/accuracy_reward/mean": 0.04233871027827263, "rewards/accuracy_reward/std": 0.2015640139579773, "step": 884 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.716796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2008.0, "completions/mean_length": 1699.19921875, "completions/mean_terminated_length": 816.3724365234375, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.30212511735085773, "frac_reward_zero_std": 0.75, "grad_norm": 7.254061258358082, "learning_rate": 8.929908759980467e-07, "loss": 0.017, "num_tokens": 861372335.0, "reward": 0.041015625, "reward_std": 0.0854918509721756, "rewards/accuracy_reward/mean": 0.04233871027827263, "rewards/accuracy_reward/std": 0.2015640139579773, "step": 885 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.806640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2008.0, "completions/mean_length": 1808.37890625, "completions/mean_terminated_length": 808.7474975585938, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.30246650166424854, "frac_reward_zero_std": 0.78125, "grad_norm": 13.18841138526206, "learning_rate": 8.926435889117019e-07, "loss": 0.0145, "num_tokens": 862373649.0, "reward": 0.037109375, "reward_std": 0.07800643146038055, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 886 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.83203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1970.0, "completions/mean_length": 1848.21484375, "completions/mean_terminated_length": 859.0348510742188, "completions/min_length": 8.0, "completions/min_terminated_length": 8.0, "epoch": 0.3028078859776393, "frac_reward_zero_std": 0.84375, "grad_norm": 13.893986923322798, "learning_rate": 8.922958155054527e-07, "loss": 0.0089, "num_tokens": 863406831.0, "reward": 0.041015625, "reward_std": 0.05814187228679657, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 887 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.751953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2007.0, "completions/mean_length": 1737.130859375, "completions/mean_terminated_length": 794.7322998046875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3031492702910301, "frac_reward_zero_std": 0.78125, "grad_norm": 10.029818981802405, "learning_rate": 8.919475562729004e-07, "loss": 0.0114, "num_tokens": 864381922.0, "reward": 0.02734375, "reward_std": 0.06987475603818893, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 888 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.814453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2030.0, "completions/mean_length": 1820.1171875, "completions/mean_terminated_length": 819.8316040039062, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.30349065460442093, "frac_reward_zero_std": 0.84375, "grad_norm": 16.276006276442953, "learning_rate": 8.915988117083351e-07, "loss": -0.0054, "num_tokens": 865388206.0, "reward": 0.060546875, "reward_std": 0.06218091398477554, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.24230584502220154, "step": 889 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2039.0, "completions/mean_length": 1824.611328125, "completions/mean_terminated_length": 804.7935180664062, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.30383203891781174, "frac_reward_zero_std": 0.90625, "grad_norm": 2.8889294067874864, "learning_rate": 8.912495823067356e-07, "loss": 0.0112, "num_tokens": 866393655.0, "reward": 0.044921875, "reward_std": 0.04396171122789383, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 890 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.775390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1958.0, "completions/mean_length": 1794.994140625, "completions/mean_terminated_length": 921.5738525390625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.3041734232312025, "frac_reward_zero_std": 0.8125, "grad_norm": 6.512924424167205, "learning_rate": 8.908998685637696e-07, "loss": 0.0157, "num_tokens": 867389764.0, "reward": 0.0546875, "reward_std": 0.08394189178943634, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 891 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.83203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1932.0, "completions/mean_length": 1872.921875, "completions/mean_terminated_length": 1005.6744384765625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.3045148075445933, "frac_reward_zero_std": 0.90625, "grad_norm": 3.7624399158122857, "learning_rate": 8.905496709757917e-07, "loss": 0.0058, "num_tokens": 868415836.0, "reward": 0.025390625, "reward_std": 0.03878315910696983, "rewards/accuracy_reward/mean": 0.026209676638245583, "rewards/accuracy_reward/std": 0.1599196344614029, "step": 892 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.845703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1938.0, "completions/mean_length": 1869.724609375, "completions/mean_terminated_length": 892.594970703125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.3048561918579841, "frac_reward_zero_std": 0.90625, "grad_norm": 10.170331309747356, "learning_rate": 8.90198990039843e-07, "loss": 0.0048, "num_tokens": 869450527.0, "reward": 0.013671875, "reward_std": 0.03394509106874466, "rewards/accuracy_reward/mean": 0.013671875, "rewards/accuracy_reward/std": 0.1162383034825325, "step": 893 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.880859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1998.0, "completions/mean_length": 1910.716796875, "completions/mean_terminated_length": 895.7212524414062, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.30519757617137494, "frac_reward_zero_std": 0.875, "grad_norm": 2.842252475556774, "learning_rate": 8.898478262536513e-07, "loss": 0.0053, "num_tokens": 870510542.0, "reward": 0.01171875, "reward_std": 0.03697281330823898, "rewards/accuracy_reward/mean": 0.01171875, "rewards/accuracy_reward/std": 0.10772226005792618, "step": 894 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.814453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1987.0, "completions/mean_length": 1816.4140625, "completions/mean_terminated_length": 799.8737182617188, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3055389604847657, "frac_reward_zero_std": 0.75, "grad_norm": 7.014274401936735, "learning_rate": 8.894961801156292e-07, "loss": 0.0218, "num_tokens": 871509202.0, "reward": 0.103515625, "reward_std": 0.10843831300735474, "rewards/accuracy_reward/mean": 0.103515625, "rewards/accuracy_reward/std": 0.30492907762527466, "step": 895 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1896.0, "completions/mean_length": 1836.443359375, "completions/mean_terminated_length": 817.125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3058803447981565, "frac_reward_zero_std": 0.90625, "grad_norm": 3.434047514521299, "learning_rate": 8.891440521248742e-07, "loss": 0.0115, "num_tokens": 872522325.0, "reward": 0.017578125, "reward_std": 0.03163585811853409, "rewards/accuracy_reward/mean": 0.017578125, "rewards/accuracy_reward/std": 0.13154059648513794, "step": 896 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.783203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1963.0, "completions/mean_length": 1794.603515625, "completions/mean_terminated_length": 879.18017578125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.3062217291115473, "frac_reward_zero_std": 0.78125, "grad_norm": 17.531396339008143, "learning_rate": 8.887914427811676e-07, "loss": 0.018, "num_tokens": 873530586.0, "reward": 0.048828125, "reward_std": 0.08159180730581284, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 897 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1901.0, "completions/mean_length": 1874.28125, "completions/mean_terminated_length": 812.6666870117188, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.30656311342493814, "frac_reward_zero_std": 0.84375, "grad_norm": 5.359007267099055, "learning_rate": 8.884383525849736e-07, "loss": 0.0161, "num_tokens": 874562506.0, "reward": 0.029296875, "reward_std": 0.05688370764255524, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17417415976524353, "step": 898 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.705078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2027.0, "completions/mean_length": 1678.982421875, "completions/mean_terminated_length": 796.7615966796875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3069044977383289, "frac_reward_zero_std": 0.78125, "grad_norm": 15.517779518403353, "learning_rate": 8.880847820374395e-07, "loss": 0.0053, "num_tokens": 875497953.0, "reward": 0.0234375, "reward_std": 0.06805649399757385, "rewards/accuracy_reward/mean": 0.0234375, "rewards/accuracy_reward/std": 0.15143637359142303, "step": 899 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.794921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1967.0, "completions/mean_length": 1805.34765625, "completions/mean_terminated_length": 864.781005859375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3072458820517197, "frac_reward_zero_std": 0.90625, "grad_norm": 6.104829379012291, "learning_rate": 8.877307316403936e-07, "loss": 0.0017, "num_tokens": 876492867.0, "reward": 0.013671875, "reward_std": 0.032461829483509064, "rewards/accuracy_reward/mean": 0.013671875, "rewards/accuracy_reward/std": 0.1162383034825325, "step": 900 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2048.0, "completions/mean_length": 1794.666015625, "completions/mean_terminated_length": 929.836181640625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3075872663651105, "frac_reward_zero_std": 0.90625, "grad_norm": 6.41204495246442, "learning_rate": 8.87376201896346e-07, "loss": 0.0083, "num_tokens": 877497752.0, "reward": 0.01171875, "reward_std": 0.031083684414625168, "rewards/accuracy_reward/mean": 0.01171875, "rewards/accuracy_reward/std": 0.10772226005792618, "step": 901 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2026.0, "completions/mean_length": 1851.3359375, "completions/mean_terminated_length": 904.4091186523438, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.30792865067850134, "frac_reward_zero_std": 0.84375, "grad_norm": 8.238588126920247, "learning_rate": 8.870211933084868e-07, "loss": 0.0037, "num_tokens": 878528244.0, "reward": 0.03515625, "reward_std": 0.06327171623706818, "rewards/accuracy_reward/mean": 0.03629032149910927, "rewards/accuracy_reward/std": 0.1872003972530365, "step": 902 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.75, "completions/max_length": 2048.0, "completions/max_terminated_length": 1957.0, "completions/mean_length": 1747.658203125, "completions/mean_terminated_length": 846.6328125, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.30827003499189215, "frac_reward_zero_std": 0.8125, "grad_norm": 23.585498526051897, "learning_rate": 8.866657063806859e-07, "loss": 0.0022, "num_tokens": 879502949.0, "reward": 0.080078125, "reward_std": 0.08765621483325958, "rewards/accuracy_reward/mean": 0.080078125, "rewards/accuracy_reward/std": 0.271679550409317, "step": 903 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1953.0, "completions/mean_length": 1727.962890625, "completions/mean_terminated_length": 877.5785522460938, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3086114193052829, "frac_reward_zero_std": 0.78125, "grad_norm": 3.3473270874570704, "learning_rate": 8.863097416174916e-07, "loss": 0.0121, "num_tokens": 880469634.0, "reward": 0.037109375, "reward_std": 0.07817822694778442, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 904 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2027.0, "completions/mean_length": 1721.701171875, "completions/mean_terminated_length": 819.5808715820312, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.3089528036186737, "frac_reward_zero_std": 0.875, "grad_norm": 11.022643754273442, "learning_rate": 8.859532995241309e-07, "loss": -0.0157, "num_tokens": 881426041.0, "reward": 0.052734375, "reward_std": 0.05628519132733345, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 905 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.755859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1976.0, "completions/mean_length": 1750.8125, "completions/mean_terminated_length": 830.7200317382812, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.30929418793206453, "frac_reward_zero_std": 0.8125, "grad_norm": 6.801666732287964, "learning_rate": 8.855963806065085e-07, "loss": 0.0016, "num_tokens": 882403161.0, "reward": 0.041015625, "reward_std": 0.07339344918727875, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 906 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.88671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1993.0, "completions/mean_length": 1922.271484375, "completions/mean_terminated_length": 938.1206665039062, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.30963557224545535, "frac_reward_zero_std": 0.90625, "grad_norm": 0.5249261196057777, "learning_rate": 8.85238985371205e-07, "loss": 0.014, "num_tokens": 883465156.0, "reward": 0.0234375, "reward_std": 0.03779878467321396, "rewards/accuracy_reward/mean": 0.0234375, "rewards/accuracy_reward/std": 0.15143637359142303, "step": 907 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.77734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2039.0, "completions/mean_length": 1777.671875, "completions/mean_terminated_length": 834.2982788085938, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.3099769565588461, "frac_reward_zero_std": 0.65625, "grad_norm": 8.082493635451137, "learning_rate": 8.848811143254779e-07, "loss": 0.0202, "num_tokens": 884452364.0, "reward": 0.103515625, "reward_std": 0.11873330175876617, "rewards/accuracy_reward/mean": 0.103515625, "rewards/accuracy_reward/std": 0.30492907762527466, "step": 908 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1883.0, "completions/mean_length": 1813.486328125, "completions/mean_terminated_length": 893.47119140625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.3103183408722369, "frac_reward_zero_std": 0.84375, "grad_norm": 9.792449496332944, "learning_rate": 8.845227679772596e-07, "loss": 0.0075, "num_tokens": 885454453.0, "reward": 0.03515625, "reward_std": 0.05931950733065605, "rewards/accuracy_reward/mean": 0.03629032149910927, "rewards/accuracy_reward/std": 0.1872004121541977, "step": 909 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.779296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1936.0, "completions/mean_length": 1776.69921875, "completions/mean_terminated_length": 818.7433471679688, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.31065972518562773, "frac_reward_zero_std": 0.75, "grad_norm": 6.140876558577721, "learning_rate": 8.841639468351571e-07, "loss": 0.0313, "num_tokens": 886437819.0, "reward": 0.044921875, "reward_std": 0.09027662128210068, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 910 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.787109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1982.0, "completions/mean_length": 1811.90625, "completions/mean_terminated_length": 939.5320434570312, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.31100110949901855, "frac_reward_zero_std": 0.84375, "grad_norm": 5.39844096889801, "learning_rate": 8.838046514084516e-07, "loss": 0.0124, "num_tokens": 887446059.0, "reward": 0.01953125, "reward_std": 0.051493462175130844, "rewards/accuracy_reward/mean": 0.01953125, "rewards/accuracy_reward/std": 0.1385180652141571, "step": 911 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.693359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2045.0, "completions/mean_length": 1657.48046875, "completions/mean_terminated_length": 774.4586181640625, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.3113424938124093, "frac_reward_zero_std": 0.84375, "grad_norm": 13.251234577297387, "learning_rate": 8.834448822070971e-07, "loss": -0.0051, "num_tokens": 888370161.0, "reward": 0.05078125, "reward_std": 0.06085042655467987, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 912 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.751953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1965.0, "completions/mean_length": 1753.416015625, "completions/mean_terminated_length": 860.3858032226562, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3116838781258001, "frac_reward_zero_std": 0.84375, "grad_norm": 5.982841117113028, "learning_rate": 8.830846397417202e-07, "loss": 0.0175, "num_tokens": 889340806.0, "reward": 0.037109375, "reward_std": 0.06311991065740585, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 913 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.79296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2041.0, "completions/mean_length": 1822.853515625, "completions/mean_terminated_length": 960.5, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.31202526243919093, "frac_reward_zero_std": 0.96875, "grad_norm": 0.06309684439459966, "learning_rate": 8.827239245236194e-07, "loss": 0.0038, "num_tokens": 890356859.0, "reward": 0.001953125, "reward_std": 0.0078125, "rewards/accuracy_reward/mean": 0.001953125, "rewards/accuracy_reward/std": 0.04419417306780815, "step": 914 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2033.0, "completions/mean_length": 1714.5, "completions/mean_terminated_length": 878.4657592773438, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.31236664675258174, "frac_reward_zero_std": 0.84375, "grad_norm": 5.777618922363556, "learning_rate": 8.823627370647634e-07, "loss": -0.0043, "num_tokens": 891316251.0, "reward": 0.029296875, "reward_std": 0.059305962175130844, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 915 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.771484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1960.0, "completions/mean_length": 1802.12890625, "completions/mean_terminated_length": 972.4530639648438, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.3127080310659725, "frac_reward_zero_std": 0.8125, "grad_norm": 8.363956877118625, "learning_rate": 8.820010778777925e-07, "loss": 0.0044, "num_tokens": 892330269.0, "reward": 0.046875, "reward_std": 0.07696875929832458, "rewards/accuracy_reward/mean": 0.046875, "rewards/accuracy_reward/std": 0.21157780289649963, "step": 916 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.810546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1959.0, "completions/mean_length": 1813.068359375, "completions/mean_terminated_length": 807.9484252929688, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.3130494153793633, "frac_reward_zero_std": 0.90625, "grad_norm": 2.280004532631335, "learning_rate": 8.816389474760151e-07, "loss": -0.0011, "num_tokens": 893335760.0, "reward": 0.021484375, "reward_std": 0.03462383896112442, "rewards/accuracy_reward/mean": 0.02217741869390011, "rewards/accuracy_reward/std": 0.14740893244743347, "step": 917 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.783203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2045.0, "completions/mean_length": 1798.962890625, "completions/mean_terminated_length": 899.288330078125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.31339079969275413, "frac_reward_zero_std": 0.78125, "grad_norm": 14.886307869057184, "learning_rate": 8.812763463734095e-07, "loss": 0.0107, "num_tokens": 894330621.0, "reward": 0.05859375, "reward_std": 0.08804067224264145, "rewards/accuracy_reward/mean": 0.05859375, "rewards/accuracy_reward/std": 0.23509246110916138, "step": 918 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1942.0, "completions/mean_length": 1771.46484375, "completions/mean_terminated_length": 868.11669921875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.31373218400614494, "frac_reward_zero_std": 0.875, "grad_norm": 5.980282174358649, "learning_rate": 8.809132750846214e-07, "loss": 0.0139, "num_tokens": 895314587.0, "reward": 0.029296875, "reward_std": 0.04125870764255524, "rewards/accuracy_reward/mean": 0.030241934582591057, "rewards/accuracy_reward/std": 0.1714252382516861, "step": 919 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.736328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2044.0, "completions/mean_length": 1751.1015625, "completions/mean_terminated_length": 921.9851684570312, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.3140735683195357, "frac_reward_zero_std": 0.71875, "grad_norm": 8.921443965171692, "learning_rate": 8.805497341249642e-07, "loss": 0.0119, "num_tokens": 896286559.0, "reward": 0.060546875, "reward_std": 0.11014118045568466, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.24230584502220154, "step": 920 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.771484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1853.0, "completions/mean_length": 1773.103515625, "completions/mean_terminated_length": 845.0342407226562, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3144149526329265, "frac_reward_zero_std": 0.6875, "grad_norm": 9.384515976146444, "learning_rate": 8.801857240104179e-07, "loss": 0.0227, "num_tokens": 897275828.0, "reward": 0.064453125, "reward_std": 0.11093293875455856, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 921 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2014.0, "completions/mean_length": 1734.880859375, "completions/mean_terminated_length": 964.7770385742188, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3147563369463173, "frac_reward_zero_std": 0.75, "grad_norm": 9.67873326234176, "learning_rate": 8.798212452576282e-07, "loss": 0.0149, "num_tokens": 898248791.0, "reward": 0.06640625, "reward_std": 0.10594936460256577, "rewards/accuracy_reward/mean": 0.06640625, "rewards/accuracy_reward/std": 0.2492343932390213, "step": 922 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1965.0, "completions/mean_length": 1752.859375, "completions/mean_terminated_length": 788.7333984375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.31509772125970814, "frac_reward_zero_std": 0.875, "grad_norm": 7.923067328173586, "learning_rate": 8.794562983839058e-07, "loss": 0.0081, "num_tokens": 899224479.0, "reward": 0.03515625, "reward_std": 0.04599714279174805, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 923 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.787109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2012.0, "completions/mean_length": 1804.013671875, "completions/mean_terminated_length": 901.9357299804688, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.3154391055730989, "frac_reward_zero_std": 0.90625, "grad_norm": 3.6045981638347033, "learning_rate": 8.790908839072262e-07, "loss": 0.0029, "num_tokens": 900220550.0, "reward": 0.01953125, "reward_std": 0.03674772381782532, "rewards/accuracy_reward/mean": 0.01953125, "rewards/accuracy_reward/std": 0.1385180652141571, "step": 924 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.75390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1888.0, "completions/mean_length": 1758.609375, "completions/mean_terminated_length": 872.0635375976562, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3157804898864897, "frac_reward_zero_std": 0.875, "grad_norm": 2.5325988196949685, "learning_rate": 8.787250023462286e-07, "loss": -0.0003, "num_tokens": 901193886.0, "reward": 0.01171875, "reward_std": 0.036034777760505676, "rewards/accuracy_reward/mean": 0.012096773833036423, "rewards/accuracy_reward/std": 0.10942844301462173, "step": 925 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.716796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2024.0, "completions/mean_length": 1728.060546875, "completions/mean_terminated_length": 918.6344604492188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.3161218741998805, "frac_reward_zero_std": 0.84375, "grad_norm": 5.270839587951115, "learning_rate": 8.783586542202148e-07, "loss": 0.0228, "num_tokens": 902160157.0, "reward": 0.048828125, "reward_std": 0.06975477933883667, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 926 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.81640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2041.0, "completions/mean_length": 1864.259765625, "completions/mean_terminated_length": 1047.7872314453125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.31646325851327134, "frac_reward_zero_std": 0.75, "grad_norm": 14.396773222267212, "learning_rate": 8.779918400491488e-07, "loss": 0.0045, "num_tokens": 903200258.0, "reward": 0.03515625, "reward_std": 0.0868166983127594, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 927 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.779296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2000.0, "completions/mean_length": 1793.53515625, "completions/mean_terminated_length": 895.0265502929688, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3168046428266621, "frac_reward_zero_std": 0.78125, "grad_norm": 4.274668512212987, "learning_rate": 8.776245603536565e-07, "loss": 0.0219, "num_tokens": 904200644.0, "reward": 0.0625, "reward_std": 0.09777653962373734, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.2422981858253479, "step": 928 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.73046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.0, "completions/mean_length": 1743.919921875, "completions/mean_terminated_length": 919.81884765625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3171460271400529, "frac_reward_zero_std": 0.875, "grad_norm": 1.3660017585141946, "learning_rate": 8.772568156550241e-07, "loss": 0.0127, "num_tokens": 905166411.0, "reward": 0.009765625, "reward_std": 0.03411140665411949, "rewards/accuracy_reward/mean": 0.009765625, "rewards/accuracy_reward/std": 0.09843364357948303, "step": 929 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.74609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1929.0, "completions/mean_length": 1759.935546875, "completions/mean_terminated_length": 913.46923828125, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.3174874114534437, "frac_reward_zero_std": 0.875, "grad_norm": 8.23438429773587, "learning_rate": 8.76888606475198e-07, "loss": -0.0024, "num_tokens": 906143338.0, "reward": 0.041015625, "reward_std": 0.04698248207569122, "rewards/accuracy_reward/mean": 0.04233871027827263, "rewards/accuracy_reward/std": 0.2015640139579773, "step": 930 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.736328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1934.0, "completions/mean_length": 1732.560546875, "completions/mean_terminated_length": 851.6666259765625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.31782879576683454, "frac_reward_zero_std": 0.75, "grad_norm": 19.068153609397328, "learning_rate": 8.765199333367837e-07, "loss": 0.0207, "num_tokens": 907115481.0, "reward": 0.08203125, "reward_std": 0.08851956576108932, "rewards/accuracy_reward/mean": 0.08203125, "rewards/accuracy_reward/std": 0.2746807038784027, "step": 931 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.68359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2033.0, "completions/mean_length": 1677.314453125, "completions/mean_terminated_length": 876.4506225585938, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.3181701800802253, "frac_reward_zero_std": 0.78125, "grad_norm": 5.399880525501455, "learning_rate": 8.761507967630453e-07, "loss": 0.0114, "num_tokens": 908049290.0, "reward": 0.01953125, "reward_std": 0.06085042655467987, "rewards/accuracy_reward/mean": 0.01953125, "rewards/accuracy_reward/std": 0.1385180652141571, "step": 932 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.79296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1879.0, "completions/mean_length": 1788.294921875, "completions/mean_terminated_length": 793.5755004882812, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.3185115643936161, "frac_reward_zero_std": 0.78125, "grad_norm": 2.8021536527336526, "learning_rate": 8.757811972779048e-07, "loss": 0.0176, "num_tokens": 909040145.0, "reward": 0.041015625, "reward_std": 0.06519509106874466, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 933 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.78125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1951.0, "completions/mean_length": 1797.64453125, "completions/mean_terminated_length": 903.5178833007812, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.3188529487070069, "frac_reward_zero_std": 0.84375, "grad_norm": 19.25262244671632, "learning_rate": 8.754111354059409e-07, "loss": 0.006, "num_tokens": 910037563.0, "reward": 0.03515625, "reward_std": 0.057269565761089325, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 934 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1927.0, "completions/mean_length": 1749.306640625, "completions/mean_terminated_length": 773.933349609375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.31919433302039774, "frac_reward_zero_std": 0.8125, "grad_norm": 18.80430393862873, "learning_rate": 8.750406116723889e-07, "loss": 0.0031, "num_tokens": 911019272.0, "reward": 0.083984375, "reward_std": 0.08378186821937561, "rewards/accuracy_reward/mean": 0.083984375, "rewards/accuracy_reward/std": 0.2776356339454651, "step": 935 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2039.0, "completions/mean_length": 1696.8203125, "completions/mean_terminated_length": 865.0789794921875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3195357173337885, "frac_reward_zero_std": 0.71875, "grad_norm": 13.62397766449746, "learning_rate": 8.746696266031392e-07, "loss": 0.0066, "num_tokens": 911970396.0, "reward": 0.03125, "reward_std": 0.08752824366092682, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 936 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.791015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2008.0, "completions/mean_length": 1767.849609375, "completions/mean_terminated_length": 707.46728515625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3198771016471793, "frac_reward_zero_std": 0.78125, "grad_norm": 14.309142070701617, "learning_rate": 8.742981807247374e-07, "loss": 0.0038, "num_tokens": 912953887.0, "reward": 0.04296875, "reward_std": 0.08191889524459839, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 937 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.78515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2044.0, "completions/mean_length": 1799.318359375, "completions/mean_terminated_length": 890.5, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3202184859605701, "frac_reward_zero_std": 0.75, "grad_norm": 7.624426304842892, "learning_rate": 8.739262745643832e-07, "loss": 0.0219, "num_tokens": 913949666.0, "reward": 0.041015625, "reward_std": 0.08961933106184006, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 938 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2000.0, "completions/mean_length": 1819.65234375, "completions/mean_terminated_length": 965.4629516601562, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.32055987027396093, "frac_reward_zero_std": 0.875, "grad_norm": 6.455680689246552, "learning_rate": 8.735539086499291e-07, "loss": 0.0139, "num_tokens": 914955888.0, "reward": 0.021484375, "reward_std": 0.04230976849794388, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 939 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.72265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1933.0, "completions/mean_length": 1721.037109375, "completions/mean_terminated_length": 869.091552734375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3209012545873517, "frac_reward_zero_std": 0.875, "grad_norm": 0.8868910171060832, "learning_rate": 8.731810835098805e-07, "loss": 0.0125, "num_tokens": 915911107.0, "reward": 0.015625, "reward_std": 0.04175759106874466, "rewards/accuracy_reward/mean": 0.015625, "rewards/accuracy_reward/std": 0.12414088100194931, "step": 940 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.779296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1989.0, "completions/mean_length": 1800.572265625, "completions/mean_terminated_length": 926.9114990234375, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.3212426389007425, "frac_reward_zero_std": 0.78125, "grad_norm": 9.161647714948735, "learning_rate": 8.728077996733945e-07, "loss": 0.0234, "num_tokens": 916908232.0, "reward": 0.0234375, "reward_std": 0.06899453699588776, "rewards/accuracy_reward/mean": 0.0234375, "rewards/accuracy_reward/std": 0.15143637359142303, "step": 941 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.751953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2041.0, "completions/mean_length": 1763.673828125, "completions/mean_terminated_length": 902.1810913085938, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.3215840232141333, "frac_reward_zero_std": 0.8125, "grad_norm": 16.70916260713759, "learning_rate": 8.724340576702791e-07, "loss": 0.0039, "num_tokens": 917908225.0, "reward": 0.021484375, "reward_std": 0.058760736137628555, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 942 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.732421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2029.0, "completions/mean_length": 1689.02734375, "completions/mean_terminated_length": 706.4379272460938, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.32192540752752413, "frac_reward_zero_std": 0.75, "grad_norm": 22.54949352326784, "learning_rate": 8.72059858030993e-07, "loss": 0.0151, "num_tokens": 918848207.0, "reward": 0.068359375, "reward_std": 0.10530310869216919, "rewards/accuracy_reward/mean": 0.068359375, "rewards/accuracy_reward/std": 0.25260838866233826, "step": 943 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.82421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1951.0, "completions/mean_length": 1836.236328125, "completions/mean_terminated_length": 843.300048828125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.3222667918409149, "frac_reward_zero_std": 0.90625, "grad_norm": 1.4544983464932375, "learning_rate": 8.716852012866438e-07, "loss": 0.0207, "num_tokens": 919864616.0, "reward": 0.03515625, "reward_std": 0.03674772381782532, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 944 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.779296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2042.0, "completions/mean_length": 1793.953125, "completions/mean_terminated_length": 896.9203491210938, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3226081761543057, "frac_reward_zero_std": 0.84375, "grad_norm": 7.848257483305533, "learning_rate": 8.713100879689886e-07, "loss": -0.0011, "num_tokens": 920858048.0, "reward": 0.037109375, "reward_std": 0.06311991065740585, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 945 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.771484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2018.0, "completions/mean_length": 1779.28515625, "completions/mean_terminated_length": 872.4701538085938, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.3229495604676965, "frac_reward_zero_std": 0.9375, "grad_norm": 0.2865516108615734, "learning_rate": 8.709345186104319e-07, "loss": 0.0114, "num_tokens": 921855746.0, "reward": 0.00390625, "reward_std": 0.015625, "rewards/accuracy_reward/mean": 0.00390625, "rewards/accuracy_reward/std": 0.06243881583213806, "step": 946 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1991.0, "completions/mean_length": 1811.861328125, "completions/mean_terminated_length": 928.9629516601562, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.32329094478108733, "frac_reward_zero_std": 0.90625, "grad_norm": 7.67541637201727, "learning_rate": 8.705584937440257e-07, "loss": -0.0009, "num_tokens": 922861595.0, "reward": 0.025390625, "reward_std": 0.03411140665411949, "rewards/accuracy_reward/mean": 0.025390625, "rewards/accuracy_reward/std": 0.15746226906776428, "step": 947 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1821.953125, "completions/mean_terminated_length": 890.6399536132812, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3236323290944781, "frac_reward_zero_std": 0.90625, "grad_norm": 5.609864691782311, "learning_rate": 8.701820139034686e-07, "loss": 0.0005, "num_tokens": 923863651.0, "reward": 0.0390625, "reward_std": 0.04659565910696983, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 948 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.818359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1814.0, "completions/mean_length": 1832.80859375, "completions/mean_terminated_length": 863.2903442382812, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3239737134078689, "frac_reward_zero_std": 0.84375, "grad_norm": 5.2184090147302316, "learning_rate": 8.698050796231049e-07, "loss": 0.0025, "num_tokens": 924878417.0, "reward": 0.03515625, "reward_std": 0.05738259106874466, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 949 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.76953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2023.0, "completions/mean_length": 1778.203125, "completions/mean_terminated_length": 877.6610107421875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3243150977212597, "frac_reward_zero_std": 0.84375, "grad_norm": 16.050515792077412, "learning_rate": 8.694276914379237e-07, "loss": 0.0159, "num_tokens": 925875129.0, "reward": 0.03125, "reward_std": 0.06194227933883667, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 950 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.84765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1640.0, "completions/mean_length": 1849.541015625, "completions/mean_terminated_length": 745.2948608398438, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.32465648203465053, "frac_reward_zero_std": 0.78125, "grad_norm": 21.705077755615644, "learning_rate": 8.690498498835586e-07, "loss": 0.0035, "num_tokens": 926906014.0, "reward": 0.044921875, "reward_std": 0.07806520164012909, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 951 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.822265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2036.0, "completions/mean_length": 1834.107421875, "completions/mean_terminated_length": 844.5604858398438, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.3249978663480413, "frac_reward_zero_std": 0.90625, "grad_norm": 2.828066311661244, "learning_rate": 8.686715554962869e-07, "loss": 0.003, "num_tokens": 927924613.0, "reward": 0.02734375, "reward_std": 0.036034777760505676, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 952 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.833984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2037.0, "completions/mean_length": 1832.306640625, "completions/mean_terminated_length": 748.7647094726562, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.3253392506614321, "frac_reward_zero_std": 0.8125, "grad_norm": 4.047753289758267, "learning_rate": 8.682928088130278e-07, "loss": 0.0286, "num_tokens": 928942610.0, "reward": 0.04296875, "reward_std": 0.07135801762342453, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 953 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.751953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2021.0, "completions/mean_length": 1769.802734375, "completions/mean_terminated_length": 926.4487915039062, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3256806349748229, "frac_reward_zero_std": 0.84375, "grad_norm": 12.661853735117854, "learning_rate": 8.679136103713431e-07, "loss": 0.0046, "num_tokens": 929922669.0, "reward": 0.037109375, "reward_std": 0.06716620922088623, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 954 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.759765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2043.0, "completions/mean_length": 1737.828125, "completions/mean_terminated_length": 756.8779907226562, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3260220192882137, "frac_reward_zero_std": 0.65625, "grad_norm": 34.318837463023065, "learning_rate": 8.675339607094356e-07, "loss": 0.0051, "num_tokens": 930885333.0, "reward": 0.06640625, "reward_std": 0.1088760495185852, "rewards/accuracy_reward/mean": 0.06640625, "rewards/accuracy_reward/std": 0.2492343932390213, "step": 955 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.822265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1855.0, "completions/mean_length": 1853.853515625, "completions/mean_terminated_length": 955.6593627929688, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.3263634036016045, "frac_reward_zero_std": 0.84375, "grad_norm": 11.75063572732186, "learning_rate": 8.671538603661489e-07, "loss": 0.0059, "num_tokens": 931907882.0, "reward": 0.03125, "reward_std": 0.06100328266620636, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 956 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2022.0, "completions/mean_length": 1748.326171875, "completions/mean_terminated_length": 885.6287841796875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.3267047879149953, "frac_reward_zero_std": 0.78125, "grad_norm": 9.20335540580923, "learning_rate": 8.667733098809655e-07, "loss": 0.0337, "num_tokens": 932880241.0, "reward": 0.060546875, "reward_std": 0.09821805357933044, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.24230584502220154, "step": 957 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.802734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2025.0, "completions/mean_length": 1825.615234375, "completions/mean_terminated_length": 920.663330078125, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.3270461722283861, "frac_reward_zero_std": 0.96875, "grad_norm": 20.869769792517626, "learning_rate": 8.663923097940072e-07, "loss": -0.0014, "num_tokens": 933890684.0, "reward": 0.001953125, "reward_std": 0.0078125, "rewards/accuracy_reward/mean": 0.002016128972172737, "rewards/accuracy_reward/std": 0.044901326298713684, "step": 958 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.806640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1974.0, "completions/mean_length": 1805.70703125, "completions/mean_terminated_length": 794.9293212890625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3273875565417769, "frac_reward_zero_std": 0.875, "grad_norm": 10.414152253488377, "learning_rate": 8.660108606460343e-07, "loss": 0.009, "num_tokens": 934887862.0, "reward": 0.048828125, "reward_std": 0.05190078169107437, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 959 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.78515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1977.0, "completions/mean_length": 1791.150390625, "completions/mean_terminated_length": 852.4818115234375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3277289408551677, "frac_reward_zero_std": 0.71875, "grad_norm": 16.301711215221964, "learning_rate": 8.656289629784439e-07, "loss": 0.0187, "num_tokens": 935882051.0, "reward": 0.0546875, "reward_std": 0.10064489394426346, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 960 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.775390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2027.0, "completions/mean_length": 1795.9609375, "completions/mean_terminated_length": 925.8782348632812, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.3280703251685585, "frac_reward_zero_std": 0.8125, "grad_norm": 21.74255026428469, "learning_rate": 8.652466173332698e-07, "loss": 0.0104, "num_tokens": 936880191.0, "reward": 0.044921875, "reward_std": 0.07316835969686508, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 961 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2028.0, "completions/mean_length": 1845.66796875, "completions/mean_terminated_length": 1051.9039306640625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.3284117094819493, "frac_reward_zero_std": 0.84375, "grad_norm": 7.642718965075587, "learning_rate": 8.648638242531817e-07, "loss": 0.0086, "num_tokens": 937907445.0, "reward": 0.013671875, "reward_std": 0.04478531330823898, "rewards/accuracy_reward/mean": 0.01411290280520916, "rewards/accuracy_reward/std": 0.11807556450366974, "step": 962 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.767578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2044.0, "completions/mean_length": 1789.095703125, "completions/mean_terminated_length": 934.0588989257812, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3287530937953401, "frac_reward_zero_std": 0.8125, "grad_norm": 9.661024603204702, "learning_rate": 8.644805842814846e-07, "loss": 0.0121, "num_tokens": 938906054.0, "reward": 0.029296875, "reward_std": 0.064970001578331, "rewards/accuracy_reward/mean": 0.030241934582591057, "rewards/accuracy_reward/std": 0.1714252382516861, "step": 963 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.80078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1962.0, "completions/mean_length": 1837.978515625, "completions/mean_terminated_length": 993.7745361328125, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.3290944781087309, "frac_reward_zero_std": 0.78125, "grad_norm": 3.5101179323687752, "learning_rate": 8.640968979621174e-07, "loss": 0.0197, "num_tokens": 939924475.0, "reward": 0.037109375, "reward_std": 0.07638143002986908, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 964 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.75, "completions/max_length": 2048.0, "completions/max_terminated_length": 2018.0, "completions/mean_length": 1751.3671875, "completions/mean_terminated_length": 861.46875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.3294358624221217, "frac_reward_zero_std": 0.9375, "grad_norm": 0.29810092832554086, "learning_rate": 8.637127658396526e-07, "loss": 0.0042, "num_tokens": 940899591.0, "reward": 0.013671875, "reward_std": 0.0234375, "rewards/accuracy_reward/mean": 0.013671875, "rewards/accuracy_reward/std": 0.1162383034825325, "step": 965 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.791015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2015.0, "completions/mean_length": 1801.28125, "completions/mean_terminated_length": 867.9345703125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3297772467355125, "frac_reward_zero_std": 0.75, "grad_norm": 10.962705122686955, "learning_rate": 8.633281884592957e-07, "loss": 0.0161, "num_tokens": 941912167.0, "reward": 0.025390625, "reward_std": 0.07438573986291885, "rewards/accuracy_reward/mean": 0.025390625, "rewards/accuracy_reward/std": 0.15746226906776428, "step": 966 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.80859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1917.0, "completions/mean_length": 1819.890625, "completions/mean_terminated_length": 856.244873046875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.3301186310489033, "frac_reward_zero_std": 0.84375, "grad_norm": 2.8134596153091294, "learning_rate": 8.629431663668834e-07, "loss": 0.0041, "num_tokens": 942921999.0, "reward": 0.01953125, "reward_std": 0.04907120764255524, "rewards/accuracy_reward/mean": 0.02016128972172737, "rewards/accuracy_reward/std": 0.14069372415542603, "step": 967 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.79296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2003.0, "completions/mean_length": 1803.453125, "completions/mean_terminated_length": 866.79248046875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3304600153622941, "frac_reward_zero_std": 0.8125, "grad_norm": 2.4968572175033015, "learning_rate": 8.625577001088848e-07, "loss": 0.0121, "num_tokens": 943919511.0, "reward": 0.056640625, "reward_std": 0.06041031330823898, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 968 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.708984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1978.0, "completions/mean_length": 1701.3984375, "completions/mean_terminated_length": 856.9932861328125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3308013996756849, "frac_reward_zero_std": 0.8125, "grad_norm": 9.808227081900512, "learning_rate": 8.621717902323987e-07, "loss": 0.0105, "num_tokens": 944865619.0, "reward": 0.03125, "reward_std": 0.06657323986291885, "rewards/accuracy_reward/mean": 0.032258063554763794, "rewards/accuracy_reward/std": 0.17686307430267334, "step": 969 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.76953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1964.0, "completions/mean_length": 1769.908203125, "completions/mean_terminated_length": 841.3643798828125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3311427839890757, "frac_reward_zero_std": 0.75, "grad_norm": 3.4634339860374457, "learning_rate": 8.61785437285153e-07, "loss": 0.0229, "num_tokens": 945853156.0, "reward": 0.037109375, "reward_std": 0.08444078266620636, "rewards/accuracy_reward/mean": 0.038306452333927155, "rewards/accuracy_reward/std": 0.19212883710861206, "step": 970 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.755859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1970.0, "completions/mean_length": 1761.810546875, "completions/mean_terminated_length": 876.4400634765625, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.3314841683024665, "frac_reward_zero_std": 0.8125, "grad_norm": 21.771348751685164, "learning_rate": 8.613986418155055e-07, "loss": 0.005, "num_tokens": 946846355.0, "reward": 0.0625, "reward_std": 0.0737205371260643, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.2422981858253479, "step": 971 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.70703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1992.0, "completions/mean_length": 1749.56640625, "completions/mean_terminated_length": 1029.3466796875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3318255526158573, "frac_reward_zero_std": 0.71875, "grad_norm": 7.6395987660404225, "learning_rate": 8.610114043724416e-07, "loss": 0.0137, "num_tokens": 947817189.0, "reward": 0.0390625, "reward_std": 0.0876617580652237, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 972 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2037.0, "completions/mean_length": 1797.380859375, "completions/mean_terminated_length": 764.8299560546875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.3321669369292481, "frac_reward_zero_std": 0.71875, "grad_norm": 15.719642914726535, "learning_rate": 8.606237255055738e-07, "loss": 0.0073, "num_tokens": 948815688.0, "reward": 0.05859375, "reward_std": 0.10542967915534973, "rewards/accuracy_reward/mean": 0.05859375, "rewards/accuracy_reward/std": 0.23509246110916138, "step": 973 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.783203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2040.0, "completions/mean_length": 1813.359375, "completions/mean_terminated_length": 965.6937255859375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.3325083212426389, "frac_reward_zero_std": 0.9375, "grad_norm": 4.4611792070603915, "learning_rate": 8.602356057651416e-07, "loss": -0.0008, "num_tokens": 949827776.0, "reward": 0.01171875, "reward_std": 0.024649331346154213, "rewards/accuracy_reward/mean": 0.01171875, "rewards/accuracy_reward/std": 0.10772226005792618, "step": 974 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.798828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2025.0, "completions/mean_length": 1823.427734375, "completions/mean_terminated_length": 931.6796264648438, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3328497055560297, "frac_reward_zero_std": 0.875, "grad_norm": 7.819143888207121, "learning_rate": 8.598470457020101e-07, "loss": 0.0062, "num_tokens": 950834955.0, "reward": 0.03125, "reward_std": 0.05259781330823898, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 975 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.771484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1968.0, "completions/mean_length": 1766.220703125, "completions/mean_terminated_length": 814.91455078125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3331910898694205, "frac_reward_zero_std": 0.78125, "grad_norm": 10.26958225534278, "learning_rate": 8.594580458676688e-07, "loss": 0.0188, "num_tokens": 951819724.0, "reward": 0.048828125, "reward_std": 0.07476464658975601, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 976 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1957.0, "completions/mean_length": 1739.337890625, "completions/mean_terminated_length": 850.7651977539062, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3335324741828113, "frac_reward_zero_std": 0.75, "grad_norm": 6.037389837638769, "learning_rate": 8.59068606814232e-07, "loss": 0.0077, "num_tokens": 952782473.0, "reward": 0.095703125, "reward_std": 0.10712699592113495, "rewards/accuracy_reward/mean": 0.095703125, "rewards/accuracy_reward/std": 0.2944713830947876, "step": 977 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.73046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2012.0, "completions/mean_length": 1718.36328125, "completions/mean_terminated_length": 825.0, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3338738584962021, "frac_reward_zero_std": 0.84375, "grad_norm": 3.647603936830429, "learning_rate": 8.586787290944373e-07, "loss": -0.0059, "num_tokens": 953742419.0, "reward": 0.048828125, "reward_std": 0.0659080371260643, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 978 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.669921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1980.0, "completions/mean_length": 1683.373046875, "completions/mean_terminated_length": 943.3313598632812, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3342152428095929, "frac_reward_zero_std": 0.71875, "grad_norm": 7.306868722803917, "learning_rate": 8.582884132616448e-07, "loss": 0.0175, "num_tokens": 954675378.0, "reward": 0.0703125, "reward_std": 0.10821780562400818, "rewards/accuracy_reward/mean": 0.0703125, "rewards/accuracy_reward/std": 0.25592297315597534, "step": 979 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.748046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1902.0, "completions/mean_length": 1760.595703125, "completions/mean_terminated_length": 907.2945556640625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3345566271229837, "frac_reward_zero_std": 0.90625, "grad_norm": 4.699939905809723, "learning_rate": 8.578976598698364e-07, "loss": 0.0112, "num_tokens": 955648659.0, "reward": 0.037109375, "reward_std": 0.03957492858171463, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 980 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.728515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1987.0, "completions/mean_length": 1726.98828125, "completions/mean_terminated_length": 865.568359375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3348980114363745, "frac_reward_zero_std": 0.75, "grad_norm": 5.5182520467823, "learning_rate": 8.57506469473615e-07, "loss": 0.0248, "num_tokens": 956606957.0, "reward": 0.060546875, "reward_std": 0.0967666283249855, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 981 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.712890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1816.0, "completions/mean_length": 1697.81640625, "completions/mean_terminated_length": 828.3129272460938, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3352393957497653, "frac_reward_zero_std": 0.875, "grad_norm": 3.042144755118716, "learning_rate": 8.57114842628204e-07, "loss": 0.0035, "num_tokens": 957553983.0, "reward": 0.03515625, "reward_std": 0.04505911096930504, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 982 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.66015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2007.0, "completions/mean_length": 1669.677734375, "completions/mean_terminated_length": 934.77587890625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.3355807800631561, "frac_reward_zero_std": 0.625, "grad_norm": 4.275034946582434, "learning_rate": 8.567227798894458e-07, "loss": 0.0237, "num_tokens": 958495482.0, "reward": 0.078125, "reward_std": 0.1371595412492752, "rewards/accuracy_reward/mean": 0.078125, "rewards/accuracy_reward/std": 0.26863065361976624, "step": 983 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.75, "completions/max_length": 2048.0, "completions/max_terminated_length": 2000.0, "completions/mean_length": 1717.796875, "completions/mean_terminated_length": 727.1875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.3359221643765469, "frac_reward_zero_std": 0.6875, "grad_norm": 14.239763764517782, "learning_rate": 8.56330281813802e-07, "loss": 0.0289, "num_tokens": 959451954.0, "reward": 0.1015625, "reward_std": 0.10707925260066986, "rewards/accuracy_reward/mean": 0.1015625, "rewards/accuracy_reward/std": 0.30236753821372986, "step": 984 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.76953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1988.0, "completions/mean_length": 1777.353515625, "completions/mean_terminated_length": 873.6694946289062, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.3362635486899377, "frac_reward_zero_std": 0.75, "grad_norm": 5.742336370284916, "learning_rate": 8.559373489583518e-07, "loss": 0.0023, "num_tokens": 960450791.0, "reward": 0.052734375, "reward_std": 0.09303291887044907, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 985 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.75, "completions/max_length": 2048.0, "completions/max_terminated_length": 2031.0, "completions/mean_length": 1759.09375, "completions/mean_terminated_length": 892.375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.3366049330033285, "frac_reward_zero_std": 0.84375, "grad_norm": 4.5892656054758385, "learning_rate": 8.555439818807914e-07, "loss": 0.0131, "num_tokens": 961432663.0, "reward": 0.017578125, "reward_std": 0.048086829483509064, "rewards/accuracy_reward/mean": 0.017578125, "rewards/accuracy_reward/std": 0.13154059648513794, "step": 986 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.740234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2008.0, "completions/mean_length": 1761.66015625, "completions/mean_terminated_length": 945.6992797851562, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3369463173167193, "frac_reward_zero_std": 0.875, "grad_norm": 5.542126449219264, "learning_rate": 8.551501811394335e-07, "loss": 0.001, "num_tokens": 962414057.0, "reward": 0.02734375, "reward_std": 0.03839729726314545, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 987 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.748046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2027.0, "completions/mean_length": 1771.013671875, "completions/mean_terminated_length": 948.6434326171875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.33728770163011007, "frac_reward_zero_std": 0.8125, "grad_norm": 1.0654922763405843, "learning_rate": 8.547559472932062e-07, "loss": 0.0158, "num_tokens": 963397616.0, "reward": 0.033203125, "reward_std": 0.06508206576108932, "rewards/accuracy_reward/mean": 0.033203125, "rewards/accuracy_reward/std": 0.17934183776378632, "step": 988 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.728515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1984.0, "completions/mean_length": 1739.828125, "completions/mean_terminated_length": 912.8633422851562, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.3376290859435009, "frac_reward_zero_std": 0.71875, "grad_norm": 9.357957290227729, "learning_rate": 8.543612809016524e-07, "loss": 0.0249, "num_tokens": 964369928.0, "reward": 0.0546875, "reward_std": 0.09640534222126007, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 989 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.76953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1897.0, "completions/mean_length": 1761.2734375, "completions/mean_terminated_length": 803.8983154296875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.3379704702568917, "frac_reward_zero_std": 0.875, "grad_norm": 6.825673397274351, "learning_rate": 8.539661825249287e-07, "loss": 0.0092, "num_tokens": 965353892.0, "reward": 0.0234375, "reward_std": 0.04670868441462517, "rewards/accuracy_reward/mean": 0.0234375, "rewards/accuracy_reward/std": 0.15143637359142303, "step": 990 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.705078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2040.0, "completions/mean_length": 1690.29296875, "completions/mean_terminated_length": 835.1126098632812, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.3383118545702825, "frac_reward_zero_std": 0.8125, "grad_norm": 4.0345676063068305, "learning_rate": 8.535706527238051e-07, "loss": 0.0206, "num_tokens": 966302026.0, "reward": 0.033203125, "reward_std": 0.06997986882925034, "rewards/accuracy_reward/mean": 0.03427419438958168, "rewards/accuracy_reward/std": 0.18211629986763, "step": 991 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.755859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1995.0, "completions/mean_length": 1737.9296875, "completions/mean_terminated_length": 777.9520263671875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.33865323888367327, "frac_reward_zero_std": 0.90625, "grad_norm": 9.934603611047853, "learning_rate": 8.531746920596639e-07, "loss": 0.0026, "num_tokens": 967274150.0, "reward": 0.029296875, "reward_std": 0.03741292655467987, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 992 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.853515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2012.0, "completions/mean_length": 1887.3203125, "completions/mean_terminated_length": 951.0933837890625, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.3389946231970641, "frac_reward_zero_std": 0.9375, "grad_norm": 1.0186849558101678, "learning_rate": 8.527783010944986e-07, "loss": 0.007, "num_tokens": 968318346.0, "reward": 0.00390625, "reward_std": 0.015625, "rewards/accuracy_reward/mean": 0.00390625, "rewards/accuracy_reward/std": 0.06243881583213806, "step": 993 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1963.0, "completions/mean_length": 1744.912109375, "completions/mean_terminated_length": 970.3541870117188, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.3393360075104549, "frac_reward_zero_std": 0.78125, "grad_norm": 12.36077386710483, "learning_rate": 8.523814803909137e-07, "loss": 0.0063, "num_tokens": 969288285.0, "reward": 0.046875, "reward_std": 0.08351518213748932, "rewards/accuracy_reward/mean": 0.046875, "rewards/accuracy_reward/std": 0.21157780289649963, "step": 994 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.775390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1902.0, "completions/mean_length": 1774.712890625, "completions/mean_terminated_length": 831.2781982421875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.3396773918238457, "frac_reward_zero_std": 0.6875, "grad_norm": 9.363849066087976, "learning_rate": 8.51984230512124e-07, "loss": 0.0113, "num_tokens": 970273754.0, "reward": 0.048828125, "reward_std": 0.1028466522693634, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 995 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.76953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1926.0, "completions/mean_length": 1806.115234375, "completions/mean_terminated_length": 998.4661254882812, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.34001877613723647, "frac_reward_zero_std": 0.78125, "grad_norm": 2.830418117651566, "learning_rate": 8.515865520219526e-07, "loss": 0.0191, "num_tokens": 971273349.0, "reward": 0.03515625, "reward_std": 0.0788990929722786, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 996 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.818359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2030.0, "completions/mean_length": 1851.8515625, "completions/mean_terminated_length": 968.1290283203125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3403601604506273, "frac_reward_zero_std": 0.8125, "grad_norm": 12.016076237145231, "learning_rate": 8.511884454848315e-07, "loss": 0.0175, "num_tokens": 972300217.0, "reward": 0.041015625, "reward_std": 0.07527707517147064, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 997 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.779296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.0, "completions/mean_length": 1776.607421875, "completions/mean_terminated_length": 818.3274536132812, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.3407015447640181, "frac_reward_zero_std": 0.78125, "grad_norm": 6.768527857697114, "learning_rate": 8.507899114658003e-07, "loss": 0.0138, "num_tokens": 973289824.0, "reward": 0.05078125, "reward_std": 0.077679343521595, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 998 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.83203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1958.0, "completions/mean_length": 1868.08984375, "completions/mean_terminated_length": 976.906982421875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.3410429290774089, "frac_reward_zero_std": 0.8125, "grad_norm": 4.054774148068096, "learning_rate": 8.503909505305048e-07, "loss": 0.024, "num_tokens": 974318286.0, "reward": 0.025390625, "reward_std": 0.06041031330823898, "rewards/accuracy_reward/mean": 0.025390625, "rewards/accuracy_reward/std": 0.15746226906776428, "step": 999 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.791015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1964.0, "completions/mean_length": 1803.578125, "completions/mean_terminated_length": 878.4298706054688, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.34138431339079967, "frac_reward_zero_std": 0.9375, "grad_norm": 1.3049787000461057, "learning_rate": 8.499915632451975e-07, "loss": 0.0093, "num_tokens": 975320662.0, "reward": 0.052734375, "reward_std": 0.0234375, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 1000 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.740234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1849.0, "completions/mean_length": 1757.51171875, "completions/mean_terminated_length": 930.0300903320312, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.3417256977041905, "frac_reward_zero_std": 0.59375, "grad_norm": 13.930003618726023, "learning_rate": 8.495917501767352e-07, "loss": 0.0377, "num_tokens": 976304732.0, "reward": 0.08203125, "reward_std": 0.15245094895362854, "rewards/accuracy_reward/mean": 0.08203125, "rewards/accuracy_reward/std": 0.2746807038784027, "step": 1001 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1962.0, "completions/mean_length": 1848.224609375, "completions/mean_terminated_length": 982.53125, "completions/min_length": 10.0, "completions/min_terminated_length": 10.0, "epoch": 0.3420670820175813, "frac_reward_zero_std": 0.875, "grad_norm": 2.9165513512615133, "learning_rate": 8.491915118925798e-07, "loss": 0.0176, "num_tokens": 977332943.0, "reward": 0.021484375, "reward_std": 0.043847277760505676, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 1002 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.69140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2030.0, "completions/mean_length": 1691.412109375, "completions/mean_terminated_length": 892.7405395507812, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3424084663309721, "frac_reward_zero_std": 0.84375, "grad_norm": 5.944478263852318, "learning_rate": 8.48790848960796e-07, "loss": 0.0041, "num_tokens": 978278930.0, "reward": 0.017578125, "reward_std": 0.048086829483509064, "rewards/accuracy_reward/mean": 0.017578125, "rewards/accuracy_reward/std": 0.13154059648513794, "step": 1003 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2032.0, "completions/mean_length": 1812.060546875, "completions/mean_terminated_length": 1006.612060546875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.3427498506443629, "frac_reward_zero_std": 0.875, "grad_norm": 4.585752433978029, "learning_rate": 8.483897619500517e-07, "loss": 0.008, "num_tokens": 979283553.0, "reward": 0.013671875, "reward_std": 0.03741292655467987, "rewards/accuracy_reward/mean": 0.01411290280520916, "rewards/accuracy_reward/std": 0.11807556450366974, "step": 1004 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.736328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1963.0, "completions/mean_length": 1737.5, "completions/mean_terminated_length": 870.3999633789062, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3430912349577537, "frac_reward_zero_std": 0.65625, "grad_norm": 12.393198351613488, "learning_rate": 8.479882514296165e-07, "loss": 0.0177, "num_tokens": 980253185.0, "reward": 0.078125, "reward_std": 0.1258384883403778, "rewards/accuracy_reward/mean": 0.078125, "rewards/accuracy_reward/std": 0.26863065361976624, "step": 1005 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2030.0, "completions/mean_length": 1846.626953125, "completions/mean_terminated_length": 927.3152465820312, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.3434326192711445, "frac_reward_zero_std": 0.75, "grad_norm": 9.630363797380602, "learning_rate": 8.475863179693613e-07, "loss": 0.0184, "num_tokens": 981269922.0, "reward": 0.048828125, "reward_std": 0.09271937608718872, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 1006 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2042.0, "completions/mean_length": 1860.923828125, "completions/mean_terminated_length": 959.5568237304688, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.3437740035845353, "frac_reward_zero_std": 0.8125, "grad_norm": 9.914601936552197, "learning_rate": 8.471839621397569e-07, "loss": 0.0116, "num_tokens": 982296203.0, "reward": 0.025390625, "reward_std": 0.06354551762342453, "rewards/accuracy_reward/mean": 0.025390625, "rewards/accuracy_reward/std": 0.15746226906776428, "step": 1007 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.87109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1886.0, "completions/mean_length": 1888.416015625, "completions/mean_terminated_length": 810.0151977539062, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3441153878979261, "frac_reward_zero_std": 0.90625, "grad_norm": 5.491343950432109, "learning_rate": 8.467811845118741e-07, "loss": 0.0007, "num_tokens": 983342224.0, "reward": 0.01171875, "reward_std": 0.031083684414625168, "rewards/accuracy_reward/mean": 0.01171875, "rewards/accuracy_reward/std": 0.10772226005792618, "step": 1008 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.775390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2018.0, "completions/mean_length": 1782.765625, "completions/mean_terminated_length": 867.13037109375, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.3444567722113169, "frac_reward_zero_std": 0.78125, "grad_norm": 28.083402465945653, "learning_rate": 8.463779856573819e-07, "loss": 0.0088, "num_tokens": 984341304.0, "reward": 0.037109375, "reward_std": 0.07481793314218521, "rewards/accuracy_reward/mean": 0.038306452333927155, "rewards/accuracy_reward/std": 0.19212885200977325, "step": 1009 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.771484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1835.0, "completions/mean_length": 1794.86328125, "completions/mean_terminated_length": 940.2564697265625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3447981565247077, "frac_reward_zero_std": 0.8125, "grad_norm": 4.020067901551045, "learning_rate": 8.459743661485472e-07, "loss": 0.0145, "num_tokens": 985336386.0, "reward": 0.060546875, "reward_std": 0.0765829086303711, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 1010 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.833984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2000.0, "completions/mean_length": 1870.310546875, "completions/mean_terminated_length": 977.682373046875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.3451395408380985, "frac_reward_zero_std": 0.875, "grad_norm": 9.43197460117714, "learning_rate": 8.455703265582342e-07, "loss": 0.0065, "num_tokens": 986369409.0, "reward": 0.03125, "reward_std": 0.04659565910696983, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 1011 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.81640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1810.0, "completions/mean_length": 1820.18359375, "completions/mean_terminated_length": 807.1276245117188, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3454809251514893, "frac_reward_zero_std": 0.90625, "grad_norm": 8.249044291063075, "learning_rate": 8.451658674599032e-07, "loss": 0.0067, "num_tokens": 987374831.0, "reward": 0.029296875, "reward_std": 0.039834219962358475, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 1012 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.78125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2042.0, "completions/mean_length": 1815.021484375, "completions/mean_terminated_length": 983.294677734375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.3458223094648801, "frac_reward_zero_std": 0.71875, "grad_norm": 17.579004426330865, "learning_rate": 8.447609894276102e-07, "loss": 0.015, "num_tokens": 988386522.0, "reward": 0.060546875, "reward_std": 0.10210913419723511, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 1013 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1885.0, "completions/mean_length": 1877.556640625, "completions/mean_terminated_length": 835.9583129882812, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.3461636937782709, "frac_reward_zero_std": 0.78125, "grad_norm": 26.053417745139626, "learning_rate": 8.443556930360048e-07, "loss": 0.0027, "num_tokens": 989417863.0, "reward": 0.060546875, "reward_std": 0.07489816844463348, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 1014 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.81640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2015.0, "completions/mean_length": 1831.9296875, "completions/mean_terminated_length": 871.1063842773438, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.3465050780916617, "frac_reward_zero_std": 0.9375, "grad_norm": 4.403770461980555, "learning_rate": 8.439499788603318e-07, "loss": -0.0021, "num_tokens": 990432211.0, "reward": 0.02734375, "reward_std": 0.024649331346154213, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 1015 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1958.0, "completions/mean_length": 1851.427734375, "completions/mean_terminated_length": 904.3068237304688, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.3468464624050525, "frac_reward_zero_std": 0.84375, "grad_norm": 8.093197422519866, "learning_rate": 8.435438474764281e-07, "loss": 0.0165, "num_tokens": 991453326.0, "reward": 0.0546875, "reward_std": 0.06535585969686508, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 1016 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.77734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1962.0, "completions/mean_length": 1796.919921875, "completions/mean_terminated_length": 920.8157958984375, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.34718784671844327, "frac_reward_zero_std": 0.84375, "grad_norm": 8.512560936995271, "learning_rate": 8.431372994607225e-07, "loss": 0.0056, "num_tokens": 992453989.0, "reward": 0.029296875, "reward_std": 0.060957908630371094, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 1017 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1922.0, "completions/mean_length": 1725.11328125, "completions/mean_terminated_length": 867.1571655273438, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3475292310318341, "frac_reward_zero_std": 0.84375, "grad_norm": 1.717356752182565, "learning_rate": 8.427303353902359e-07, "loss": 0.0178, "num_tokens": 993420751.0, "reward": 0.037109375, "reward_std": 0.05782270431518555, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 1018 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.779296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1933.0, "completions/mean_length": 1805.65625, "completions/mean_terminated_length": 949.9468994140625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.3478706153452249, "frac_reward_zero_std": 0.90625, "grad_norm": 21.660366778558416, "learning_rate": 8.423229558425796e-07, "loss": 0.0154, "num_tokens": 994416431.0, "reward": 0.021484375, "reward_std": 0.03462383896112442, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 1019 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2027.0, "completions/mean_length": 1827.24609375, "completions/mean_terminated_length": 917.739990234375, "completions/min_length": 216.0, "completions/min_terminated_length": 216.0, "epoch": 0.3482119996586157, "frac_reward_zero_std": 0.6875, "grad_norm": 9.878856782670251, "learning_rate": 8.419151613959539e-07, "loss": 0.0228, "num_tokens": 995426941.0, "reward": 0.07421875, "reward_std": 0.1165291890501976, "rewards/accuracy_reward/mean": 0.07421875, "rewards/accuracy_reward/std": 0.2623828947544098, "step": 1020 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.826171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2024.0, "completions/mean_length": 1829.515625, "completions/mean_terminated_length": 791.1011352539062, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.34855338397200647, "frac_reward_zero_std": 0.90625, "grad_norm": 2.016803461282072, "learning_rate": 8.415069526291486e-07, "loss": 0.0057, "num_tokens": 996432741.0, "reward": 0.041015625, "reward_std": 0.029160313308238983, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 1021 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2007.0, "completions/mean_length": 1799.314453125, "completions/mean_terminated_length": 1021.1693115234375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.3488947682853973, "frac_reward_zero_std": 0.875, "grad_norm": 4.893572373632423, "learning_rate": 8.410983301215415e-07, "loss": 0.0014, "num_tokens": 997426758.0, "reward": 0.017578125, "reward_std": 0.043680962175130844, "rewards/accuracy_reward/mean": 0.017578125, "rewards/accuracy_reward/std": 0.13154059648513794, "step": 1022 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.76171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1935.0, "completions/mean_length": 1788.064453125, "completions/mean_terminated_length": 957.1229248046875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.3492361525987881, "frac_reward_zero_std": 0.8125, "grad_norm": 4.75299649504567, "learning_rate": 8.406892944530973e-07, "loss": 0.0143, "num_tokens": 998417559.0, "reward": 0.0625, "reward_std": 0.07104447484016418, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.2422981858253479, "step": 1023 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.783203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1984.0, "completions/mean_length": 1806.373046875, "completions/mean_terminated_length": 933.468505859375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3495775369121789, "frac_reward_zero_std": 0.78125, "grad_norm": 12.606571273794385, "learning_rate": 8.402798462043673e-07, "loss": 0.0225, "num_tokens": 999407926.0, "reward": 0.0703125, "reward_std": 0.08738242089748383, "rewards/accuracy_reward/mean": 0.07258064299821854, "rewards/accuracy_reward/std": 0.25970885157585144, "step": 1024 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1881.0, "completions/mean_length": 1814.607421875, "completions/mean_terminated_length": 941.5463256835938, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.34991892122556967, "frac_reward_zero_std": 0.78125, "grad_norm": 31.33963611162889, "learning_rate": 8.398699859564887e-07, "loss": 0.0258, "num_tokens": 1000417709.0, "reward": 0.0625, "reward_std": 0.0904373973608017, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.2422981858253479, "step": 1025 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.794921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2029.0, "completions/mean_length": 1797.708984375, "completions/mean_terminated_length": 827.5333862304688, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.3502603055389605, "frac_reward_zero_std": 0.84375, "grad_norm": 12.226804481190621, "learning_rate": 8.39459714291183e-07, "loss": 0.0172, "num_tokens": 1001413608.0, "reward": 0.044921875, "reward_std": 0.060412678867578506, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 1026 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.830078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2031.0, "completions/mean_length": 1864.234375, "completions/mean_terminated_length": 966.5287475585938, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.3506016898523513, "frac_reward_zero_std": 0.71875, "grad_norm": 15.229421019989417, "learning_rate": 8.390490317907557e-07, "loss": 0.0298, "num_tokens": 1002439824.0, "reward": 0.048828125, "reward_std": 0.10139855742454529, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 1027 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.783203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1963.0, "completions/mean_length": 1824.28125, "completions/mean_terminated_length": 1016.0720825195312, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.3509430741657421, "frac_reward_zero_std": 0.65625, "grad_norm": 13.031020100907192, "learning_rate": 8.386379390380956e-07, "loss": 0.0271, "num_tokens": 1003446032.0, "reward": 0.072265625, "reward_std": 0.13020557165145874, "rewards/accuracy_reward/mean": 0.072265625, "rewards/accuracy_reward/std": 0.2591804563999176, "step": 1028 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1924.0, "completions/mean_length": 1807.98828125, "completions/mean_terminated_length": 819.1399536132812, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.35128445847913287, "frac_reward_zero_std": 0.875, "grad_norm": 6.46081562829866, "learning_rate": 8.382264366166736e-07, "loss": 0.0041, "num_tokens": 1004445258.0, "reward": 0.01953125, "reward_std": 0.04192390665411949, "rewards/accuracy_reward/mean": 0.01953125, "rewards/accuracy_reward/std": 0.1385180652141571, "step": 1029 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.712890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1965.0, "completions/mean_length": 1719.58984375, "completions/mean_terminated_length": 904.4557495117188, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3516258427925237, "frac_reward_zero_std": 0.8125, "grad_norm": 15.058953496918804, "learning_rate": 8.378145251105423e-07, "loss": 0.0129, "num_tokens": 1005410072.0, "reward": 0.064453125, "reward_std": 0.08648413419723511, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 1030 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2048.0, "completions/mean_length": 1751.86328125, "completions/mean_terminated_length": 1223.9674072265625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3519672271059145, "frac_reward_zero_std": 0.84375, "grad_norm": 12.985636098147594, "learning_rate": 8.374022051043344e-07, "loss": 0.0212, "num_tokens": 1006385202.0, "reward": 0.064453125, "reward_std": 0.06602106243371964, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 1031 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.794921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2039.0, "completions/mean_length": 1815.671875, "completions/mean_terminated_length": 915.1238403320312, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3523086114193053, "frac_reward_zero_std": 0.90625, "grad_norm": 8.392169320136071, "learning_rate": 8.36989477183263e-07, "loss": 0.005, "num_tokens": 1007390538.0, "reward": 0.01171875, "reward_std": 0.02960042469203472, "rewards/accuracy_reward/mean": 0.01171875, "rewards/accuracy_reward/std": 0.10772226005792618, "step": 1032 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2016.0, "completions/mean_length": 1734.767578125, "completions/mean_terminated_length": 949.5410766601562, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.35264999573269606, "frac_reward_zero_std": 0.8125, "grad_norm": 16.945360606966492, "learning_rate": 8.365763419331199e-07, "loss": 0.0091, "num_tokens": 1008354675.0, "reward": 0.041015625, "reward_std": 0.0759630799293518, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 1033 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.794921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2027.0, "completions/mean_length": 1838.7421875, "completions/mean_terminated_length": 1027.619140625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.3529913800460869, "frac_reward_zero_std": 0.8125, "grad_norm": 6.82911314401029, "learning_rate": 8.361627999402748e-07, "loss": 0.0088, "num_tokens": 1009367663.0, "reward": 0.05859375, "reward_std": 0.07179021835327148, "rewards/accuracy_reward/mean": 0.05859375, "rewards/accuracy_reward/std": 0.23509246110916138, "step": 1034 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.79296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2042.0, "completions/mean_length": 1829.51953125, "completions/mean_terminated_length": 992.6981201171875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.3533327643594777, "frac_reward_zero_std": 0.84375, "grad_norm": 19.775242794877773, "learning_rate": 8.357488517916752e-07, "loss": 0.012, "num_tokens": 1010382905.0, "reward": 0.02734375, "reward_std": 0.05012226849794388, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 1035 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.720703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1988.0, "completions/mean_length": 1718.548828125, "completions/mean_terminated_length": 868.4265747070312, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3536741486728685, "frac_reward_zero_std": 0.8125, "grad_norm": 10.056353428423002, "learning_rate": 8.353344980748446e-07, "loss": 0.0154, "num_tokens": 1011344354.0, "reward": 0.04296875, "reward_std": 0.07383356243371964, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 1036 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.837890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1854.0, "completions/mean_length": 1857.001953125, "completions/mean_terminated_length": 869.795166015625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.35401553298625926, "frac_reward_zero_std": 0.84375, "grad_norm": 13.944942695040773, "learning_rate": 8.349197393778825e-07, "loss": 0.0075, "num_tokens": 1012369811.0, "reward": 0.025390625, "reward_std": 0.0557793527841568, "rewards/accuracy_reward/mean": 0.025390625, "rewards/accuracy_reward/std": 0.15746226906776428, "step": 1037 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.806640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1994.0, "completions/mean_length": 1817.22265625, "completions/mean_terminated_length": 854.48486328125, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.3543569172996501, "frac_reward_zero_std": 0.9375, "grad_norm": 1.519135208801546, "learning_rate": 8.345045762894628e-07, "loss": 0.0154, "num_tokens": 1013371349.0, "reward": 0.01953125, "reward_std": 0.023823359981179237, "rewards/accuracy_reward/mean": 0.01953125, "rewards/accuracy_reward/std": 0.1385180652141571, "step": 1038 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.814453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1889.0, "completions/mean_length": 1811.8671875, "completions/mean_terminated_length": 775.3684692382812, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.3546983016130409, "frac_reward_zero_std": 0.78125, "grad_norm": 4.120897787669056, "learning_rate": 8.340890093988336e-07, "loss": 0.0232, "num_tokens": 1014371665.0, "reward": 0.04296875, "reward_std": 0.0794091522693634, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 1039 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2020.0, "completions/mean_length": 1852.8515625, "completions/mean_terminated_length": 858.5238037109375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3550396859264317, "frac_reward_zero_std": 0.84375, "grad_norm": 15.888927381380547, "learning_rate": 8.336730392958163e-07, "loss": -0.0007, "num_tokens": 1015392965.0, "reward": 0.021484375, "reward_std": 0.04973640665411949, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 1040 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.87890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1965.0, "completions/mean_length": 1904.0546875, "completions/mean_terminated_length": 859.290283203125, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.35538107023982246, "frac_reward_zero_std": 0.90625, "grad_norm": 24.276854607553346, "learning_rate": 8.332566665708041e-07, "loss": 0.0042, "num_tokens": 1016450305.0, "reward": 0.025390625, "reward_std": 0.030584799125790596, "rewards/accuracy_reward/mean": 0.025390625, "rewards/accuracy_reward/std": 0.15746226906776428, "step": 1041 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.716796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1969.0, "completions/mean_length": 1709.044921875, "completions/mean_terminated_length": 851.137939453125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3557224545532133, "frac_reward_zero_std": 0.78125, "grad_norm": 18.19504078478282, "learning_rate": 8.328398918147622e-07, "loss": 0.0139, "num_tokens": 1017399624.0, "reward": 0.0625, "reward_std": 0.08675792813301086, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.2422981858253479, "step": 1042 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.78125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1987.0, "completions/mean_length": 1791.02734375, "completions/mean_terminated_length": 873.2678833007812, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3560638388666041, "frac_reward_zero_std": 0.875, "grad_norm": 11.866374901890465, "learning_rate": 8.32422715619226e-07, "loss": 0.0021, "num_tokens": 1018403350.0, "reward": 0.013671875, "reward_std": 0.03889618441462517, "rewards/accuracy_reward/mean": 0.01411290280520916, "rewards/accuracy_reward/std": 0.11807556450366974, "step": 1043 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.755859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2048.0, "completions/mean_length": 1744.076171875, "completions/mean_terminated_length": 803.1280517578125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.3564052231799949, "frac_reward_zero_std": 0.75, "grad_norm": 28.050169617909198, "learning_rate": 8.320051385763005e-07, "loss": 0.02, "num_tokens": 1019382029.0, "reward": 0.08203125, "reward_std": 0.09486784040927887, "rewards/accuracy_reward/mean": 0.08467742055654526, "rewards/accuracy_reward/std": 0.278682142496109, "step": 1044 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.794921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1883.0, "completions/mean_length": 1806.453125, "completions/mean_terminated_length": 870.1714477539062, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.35674660749338566, "frac_reward_zero_std": 0.75, "grad_norm": 12.298429803880369, "learning_rate": 8.315871612786604e-07, "loss": 0.0208, "num_tokens": 1020390677.0, "reward": 0.044921875, "reward_std": 0.0921536535024643, "rewards/accuracy_reward/mean": 0.0463709682226181, "rewards/accuracy_reward/std": 0.21049949526786804, "step": 1045 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1987.0, "completions/mean_length": 1785.47265625, "completions/mean_terminated_length": 964.01611328125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.3570879918067765, "frac_reward_zero_std": 0.875, "grad_norm": 13.35327019742105, "learning_rate": 8.311687843195481e-07, "loss": 0.0038, "num_tokens": 1021382791.0, "reward": 0.02734375, "reward_std": 0.051493462175130844, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 1046 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.810546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1871.0, "completions/mean_length": 1823.455078125, "completions/mean_terminated_length": 862.773193359375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.3574293761201673, "frac_reward_zero_std": 0.78125, "grad_norm": 11.142855779593305, "learning_rate": 8.307500082927726e-07, "loss": 0.0099, "num_tokens": 1022398016.0, "reward": 0.03125, "reward_std": 0.07300759106874466, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 1047 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.76953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1951.0, "completions/mean_length": 1761.30078125, "completions/mean_terminated_length": 804.0169677734375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3577707604335581, "frac_reward_zero_std": 0.875, "grad_norm": 24.867748822848235, "learning_rate": 8.303308337927103e-07, "loss": -0.002, "num_tokens": 1023384186.0, "reward": 0.041015625, "reward_std": 0.05099457502365112, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 1048 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.78125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2037.0, "completions/mean_length": 1797.24609375, "completions/mean_terminated_length": 902.169677734375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.35811214474694886, "frac_reward_zero_std": 0.71875, "grad_norm": 25.31652117155038, "learning_rate": 8.299112614143028e-07, "loss": 0.0187, "num_tokens": 1024385368.0, "reward": 0.087890625, "reward_std": 0.11250369995832443, "rewards/accuracy_reward/mean": 0.087890625, "rewards/accuracy_reward/std": 0.2834126651287079, "step": 1049 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.80859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2005.0, "completions/mean_length": 1822.185546875, "completions/mean_terminated_length": 868.2346801757812, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.35845352906033967, "frac_reward_zero_std": 0.96875, "grad_norm": 2.1197805731457176, "learning_rate": 8.29491291753056e-07, "loss": -0.0002, "num_tokens": 1025400103.0, "reward": 0.00390625, "reward_std": 0.010673906654119492, "rewards/accuracy_reward/mean": 0.004032257944345474, "rewards/accuracy_reward/std": 0.06343589723110199, "step": 1050 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1954.0, "completions/mean_length": 1650.38671875, "completions/mean_terminated_length": 653.630126953125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3587949133737305, "frac_reward_zero_std": 0.75, "grad_norm": 13.60842730212261, "learning_rate": 8.290709254050403e-07, "loss": 0.0175, "num_tokens": 1026319629.0, "reward": 0.072265625, "reward_std": 0.08699656277894974, "rewards/accuracy_reward/mean": 0.072265625, "rewards/accuracy_reward/std": 0.2591804563999176, "step": 1051 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1999.0, "completions/mean_length": 1785.998046875, "completions/mean_terminated_length": 758.144287109375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.3591362976871213, "frac_reward_zero_std": 0.9375, "grad_norm": 6.993621953210613, "learning_rate": 8.286501629668887e-07, "loss": -0.0032, "num_tokens": 1027307276.0, "reward": 0.005859375, "reward_std": 0.01848640665411949, "rewards/accuracy_reward/mean": 0.005859375, "rewards/accuracy_reward/std": 0.07639661431312561, "step": 1052 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.744140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1950.0, "completions/mean_length": 1756.396484375, "completions/mean_terminated_length": 908.2977294921875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.35947768200051206, "frac_reward_zero_std": 0.8125, "grad_norm": 7.217672045595186, "learning_rate": 8.282290050357966e-07, "loss": 0.0209, "num_tokens": 1028287191.0, "reward": 0.048828125, "reward_std": 0.07407219707965851, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 1053 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.80859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2022.0, "completions/mean_length": 1831.001953125, "completions/mean_terminated_length": 914.744873046875, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.35981906631390287, "frac_reward_zero_std": 0.9375, "grad_norm": 6.4682384353089875, "learning_rate": 8.278074522095207e-07, "loss": 0.0016, "num_tokens": 1029310568.0, "reward": 0.013671875, "reward_std": 0.025633705779910088, "rewards/accuracy_reward/mean": 0.013671875, "rewards/accuracy_reward/std": 0.1162383034825325, "step": 1054 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.76171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1949.0, "completions/mean_length": 1761.427734375, "completions/mean_terminated_length": 845.3359985351562, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3601604506272937, "frac_reward_zero_std": 0.84375, "grad_norm": 4.251167829416981, "learning_rate": 8.273855050863779e-07, "loss": 0.0094, "num_tokens": 1030286883.0, "reward": 0.060546875, "reward_std": 0.06602106243371964, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 1055 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.75, "completions/max_length": 2048.0, "completions/max_terminated_length": 1955.0, "completions/mean_length": 1745.0078125, "completions/mean_terminated_length": 836.03125, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.3605018349406845, "frac_reward_zero_std": 0.84375, "grad_norm": 3.5187319712984806, "learning_rate": 8.269631642652454e-07, "loss": 0.0128, "num_tokens": 1031255591.0, "reward": 0.029296875, "reward_std": 0.04973640665411949, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 1056 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.79296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1765.0, "completions/mean_length": 1819.654296875, "completions/mean_terminated_length": 945.0471801757812, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.36084321925407525, "frac_reward_zero_std": 0.8125, "grad_norm": 8.074165024012252, "learning_rate": 8.265404303455583e-07, "loss": 0.0062, "num_tokens": 1032257798.0, "reward": 0.037109375, "reward_std": 0.07273616641759872, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 1057 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.748046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1995.0, "completions/mean_length": 1734.48046875, "completions/mean_terminated_length": 803.6434326171875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.36118460356746607, "frac_reward_zero_std": 0.875, "grad_norm": 17.15321905079241, "learning_rate": 8.261173039273103e-07, "loss": 0.007, "num_tokens": 1033226108.0, "reward": 0.044921875, "reward_std": 0.03741292655467987, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 1058 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.69140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1915.0, "completions/mean_length": 1673.26171875, "completions/mean_terminated_length": 833.6582641601562, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3615259878808569, "frac_reward_zero_std": 0.71875, "grad_norm": 13.817990981812805, "learning_rate": 8.25693785611052e-07, "loss": 0.0142, "num_tokens": 1034149106.0, "reward": 0.0703125, "reward_std": 0.10986737906932831, "rewards/accuracy_reward/mean": 0.0703125, "rewards/accuracy_reward/std": 0.25592297315597534, "step": 1059 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1901.0, "completions/mean_length": 1739.341796875, "completions/mean_terminated_length": 885.99267578125, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.3618673721942477, "frac_reward_zero_std": 0.78125, "grad_norm": 10.766935989511598, "learning_rate": 8.2526987599789e-07, "loss": 0.0023, "num_tokens": 1035114897.0, "reward": 0.0546875, "reward_std": 0.08832141757011414, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 1060 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2037.0, "completions/mean_length": 1741.466796875, "completions/mean_terminated_length": 859.0227661132812, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.36220875650763845, "frac_reward_zero_std": 0.71875, "grad_norm": 8.440053427541136, "learning_rate": 8.248455756894865e-07, "loss": 0.02, "num_tokens": 1036097392.0, "reward": 0.033203125, "reward_std": 0.0879673957824707, "rewards/accuracy_reward/mean": 0.033203125, "rewards/accuracy_reward/std": 0.17934183776378632, "step": 1061 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.689453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1885.0, "completions/mean_length": 1649.724609375, "completions/mean_terminated_length": 765.5031127929688, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.36255014082102927, "frac_reward_zero_std": 0.84375, "grad_norm": 1.2529459478913803, "learning_rate": 8.244208852880583e-07, "loss": -0.0022, "num_tokens": 1037015843.0, "reward": 0.044921875, "reward_std": 0.057656385004520416, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 1062 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.751953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1951.0, "completions/mean_length": 1736.06640625, "completions/mean_terminated_length": 790.44091796875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.3628915251344201, "frac_reward_zero_std": 0.78125, "grad_norm": 12.817836090554461, "learning_rate": 8.239958053963758e-07, "loss": 0.0283, "num_tokens": 1037973045.0, "reward": 0.072265625, "reward_std": 0.09781768918037415, "rewards/accuracy_reward/mean": 0.072265625, "rewards/accuracy_reward/std": 0.2591804563999176, "step": 1063 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2018.0, "completions/mean_length": 1678.244140625, "completions/mean_terminated_length": 768.8446044921875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3632329094478109, "frac_reward_zero_std": 0.84375, "grad_norm": 36.807616073665045, "learning_rate": 8.23570336617762e-07, "loss": -0.0078, "num_tokens": 1038911970.0, "reward": 0.060546875, "reward_std": 0.04957009106874466, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 1064 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.712890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1918.0, "completions/mean_length": 1714.072265625, "completions/mean_terminated_length": 884.9319458007812, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.36357429376120165, "frac_reward_zero_std": 0.71875, "grad_norm": 23.60677450689423, "learning_rate": 8.23144479556092e-07, "loss": 0.0112, "num_tokens": 1039863703.0, "reward": 0.05859375, "reward_std": 0.09671889245510101, "rewards/accuracy_reward/mean": 0.05859375, "rewards/accuracy_reward/std": 0.23509246110916138, "step": 1065 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.654296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2045.0, "completions/mean_length": 1626.23046875, "completions/mean_terminated_length": 827.9661254882812, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.36391567807459246, "frac_reward_zero_std": 0.75, "grad_norm": 11.071367208311841, "learning_rate": 8.227182348157923e-07, "loss": 0.0297, "num_tokens": 1040768013.0, "reward": 0.111328125, "reward_std": 0.1078319400548935, "rewards/accuracy_reward/mean": 0.111328125, "rewards/accuracy_reward/std": 0.31484565138816833, "step": 1066 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.72265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1996.0, "completions/mean_length": 1709.728515625, "completions/mean_terminated_length": 828.31689453125, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.3642570623879833, "frac_reward_zero_std": 0.6875, "grad_norm": 8.703691427984685, "learning_rate": 8.222916030018389e-07, "loss": 0.0132, "num_tokens": 1041723602.0, "reward": 0.078125, "reward_std": 0.1201498806476593, "rewards/accuracy_reward/mean": 0.078125, "rewards/accuracy_reward/std": 0.26863065361976624, "step": 1067 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.650390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1955.0, "completions/mean_length": 1648.720703125, "completions/mean_terminated_length": 905.9273681640625, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.3645984467013741, "frac_reward_zero_std": 0.6875, "grad_norm": 11.497292087012905, "learning_rate": 8.21864584719758e-07, "loss": 0.0199, "num_tokens": 1042644643.0, "reward": 0.078125, "reward_std": 0.1150255799293518, "rewards/accuracy_reward/mean": 0.08064515888690948, "rewards/accuracy_reward/std": 0.2725643217563629, "step": 1068 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.708984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2000.0, "completions/mean_length": 1751.888671875, "completions/mean_terminated_length": 1030.489990234375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.36493983101476485, "frac_reward_zero_std": 0.84375, "grad_norm": 16.736426792450537, "learning_rate": 8.214371805756238e-07, "loss": 0.0015, "num_tokens": 1043622794.0, "reward": 0.0234375, "reward_std": 0.05259781330823898, "rewards/accuracy_reward/mean": 0.0234375, "rewards/accuracy_reward/std": 0.15143637359142303, "step": 1069 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.615234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2010.0, "completions/mean_length": 1566.40625, "completions/mean_terminated_length": 796.3451538085938, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.36528121532815566, "frac_reward_zero_std": 0.84375, "grad_norm": 4.550799332810863, "learning_rate": 8.210093911760582e-07, "loss": 0.007, "num_tokens": 1044503290.0, "reward": 0.048828125, "reward_std": 0.064970001578331, "rewards/accuracy_reward/mean": 0.05040322616696358, "rewards/accuracy_reward/std": 0.21899642050266266, "step": 1070 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.65625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2035.0, "completions/mean_length": 1616.333984375, "completions/mean_terminated_length": 792.2443237304688, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3656225996415465, "frac_reward_zero_std": 0.71875, "grad_norm": 4.151064253424261, "learning_rate": 8.205812171282302e-07, "loss": 0.0096, "num_tokens": 1045415365.0, "reward": 0.0703125, "reward_std": 0.10916339606046677, "rewards/accuracy_reward/mean": 0.0703125, "rewards/accuracy_reward/std": 0.25592297315597534, "step": 1071 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.646484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2006.0, "completions/mean_length": 1675.703125, "completions/mean_terminated_length": 994.8729858398438, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.3659639839549373, "frac_reward_zero_std": 0.75, "grad_norm": 18.23929482315597, "learning_rate": 8.201526590398543e-07, "loss": 0.0068, "num_tokens": 1046351053.0, "reward": 0.08203125, "reward_std": 0.10189647972583771, "rewards/accuracy_reward/mean": 0.08203125, "rewards/accuracy_reward/std": 0.2746807038784027, "step": 1072 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.572265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2000.0, "completions/mean_length": 1530.111328125, "completions/mean_terminated_length": 837.3515625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.36630536826832805, "frac_reward_zero_std": 0.78125, "grad_norm": 10.377436425086909, "learning_rate": 8.197237175191907e-07, "loss": 0.0174, "num_tokens": 1047215494.0, "reward": 0.076171875, "reward_std": 0.08384227007627487, "rewards/accuracy_reward/mean": 0.0786290317773819, "rewards/accuracy_reward/std": 0.26943066716194153, "step": 1073 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.626953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1899.0, "completions/mean_length": 1581.9609375, "completions/mean_terminated_length": 798.7225341796875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.36664675258171886, "frac_reward_zero_std": 0.65625, "grad_norm": 11.366920620380323, "learning_rate": 8.192943931750431e-07, "loss": 0.0421, "num_tokens": 1048096626.0, "reward": 0.111328125, "reward_std": 0.1449257880449295, "rewards/accuracy_reward/mean": 0.11491935700178146, "rewards/accuracy_reward/std": 0.3192465901374817, "step": 1074 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.62890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2029.0, "completions/mean_length": 1604.96875, "completions/mean_terminated_length": 854.1473999023438, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.3669881368951097, "frac_reward_zero_std": 0.625, "grad_norm": 19.296695107904313, "learning_rate": 8.188646866167591e-07, "loss": 0.0101, "num_tokens": 1048995330.0, "reward": 0.076171875, "reward_std": 0.14179153740406036, "rewards/accuracy_reward/mean": 0.076171875, "rewards/accuracy_reward/std": 0.26553234457969666, "step": 1075 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.623046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1976.0, "completions/mean_length": 1573.33984375, "completions/mean_terminated_length": 789.0673217773438, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.3673295212085005, "frac_reward_zero_std": 0.75, "grad_norm": 16.395138066910892, "learning_rate": 8.184345984542283e-07, "loss": 0.0156, "num_tokens": 1049879248.0, "reward": 0.072265625, "reward_std": 0.10074453055858612, "rewards/accuracy_reward/mean": 0.072265625, "rewards/accuracy_reward/std": 0.2591804563999176, "step": 1076 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.5546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1890.0, "completions/mean_length": 1510.125, "completions/mean_terminated_length": 840.140380859375, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.36767090552189124, "frac_reward_zero_std": 0.78125, "grad_norm": 8.247313697619292, "learning_rate": 8.180041292978826e-07, "loss": 0.0148, "num_tokens": 1050719936.0, "reward": 0.052734375, "reward_std": 0.07917051762342453, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 1077 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.61328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1851.0, "completions/mean_length": 1615.302734375, "completions/mean_terminated_length": 929.1060791015625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.36801228983528206, "frac_reward_zero_std": 0.75, "grad_norm": 9.369241000743516, "learning_rate": 8.175732797586939e-07, "loss": 0.0086, "num_tokens": 1051614299.0, "reward": 0.021484375, "reward_std": 0.07014618813991547, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 1078 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.58203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2004.0, "completions/mean_length": 1543.2109375, "completions/mean_terminated_length": 840.4625854492188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.36835367414867287, "frac_reward_zero_std": 0.78125, "grad_norm": 4.94878356587409, "learning_rate": 8.171420504481743e-07, "loss": 0.0191, "num_tokens": 1052487991.0, "reward": 0.03515625, "reward_std": 0.07352001965045929, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 1079 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.66015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2029.0, "completions/mean_length": 1680.541015625, "completions/mean_terminated_length": 966.7413940429688, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3686950584620637, "frac_reward_zero_std": 0.8125, "grad_norm": 12.238720437917998, "learning_rate": 8.167104419783753e-07, "loss": 0.0037, "num_tokens": 1053430012.0, "reward": 0.0703125, "reward_std": 0.07955637574195862, "rewards/accuracy_reward/mean": 0.0703125, "rewards/accuracy_reward/std": 0.25592297315597534, "step": 1080 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.630859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2017.0, "completions/mean_length": 1613.716796875, "completions/mean_terminated_length": 871.529052734375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.36903644277545444, "frac_reward_zero_std": 0.71875, "grad_norm": 6.95260561579441, "learning_rate": 8.162784549618855e-07, "loss": 0.0315, "num_tokens": 1054333723.0, "reward": 0.03125, "reward_std": 0.08698301017284393, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 1081 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.58984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2038.0, "completions/mean_length": 1625.12890625, "completions/mean_terminated_length": 1017.0000610351562, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.36937782708884526, "frac_reward_zero_std": 0.8125, "grad_norm": 8.096928061043071, "learning_rate": 8.158460900118321e-07, "loss": 0.007, "num_tokens": 1055236573.0, "reward": 0.029296875, "reward_std": 0.06684703379869461, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 1082 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.638671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2024.0, "completions/mean_length": 1652.919921875, "completions/mean_terminated_length": 954.5892333984375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.36971921140223607, "frac_reward_zero_std": 0.75, "grad_norm": 11.00954503058334, "learning_rate": 8.15413347741878e-07, "loss": 0.0173, "num_tokens": 1056161524.0, "reward": 0.037109375, "reward_std": 0.08560487627983093, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 1083 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.53125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1959.0, "completions/mean_length": 1511.349609375, "completions/mean_terminated_length": 903.1458740234375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.3700605957156269, "frac_reward_zero_std": 0.6875, "grad_norm": 18.22106998469532, "learning_rate": 8.149802287662214e-07, "loss": 0.0226, "num_tokens": 1057006775.0, "reward": 0.091796875, "reward_std": 0.1285618543624878, "rewards/accuracy_reward/mean": 0.091796875, "rewards/accuracy_reward/std": 0.289021372795105, "step": 1084 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.58203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1907.0, "completions/mean_length": 1553.48828125, "completions/mean_terminated_length": 864.869140625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.37040198002901764, "frac_reward_zero_std": 0.84375, "grad_norm": 5.76414460599428, "learning_rate": 8.145467336995954e-07, "loss": -0.0133, "num_tokens": 1057880225.0, "reward": 0.052734375, "reward_std": 0.04973640665411949, "rewards/accuracy_reward/mean": 0.05624999850988388, "rewards/accuracy_reward/std": 0.23064424097537994, "step": 1085 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.61328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2034.0, "completions/mean_length": 1625.099609375, "completions/mean_terminated_length": 954.4393920898438, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.37074336434240845, "frac_reward_zero_std": 0.75, "grad_norm": 2.0411881148299713, "learning_rate": 8.141128631572676e-07, "loss": 0.008, "num_tokens": 1058790436.0, "reward": 0.064453125, "reward_std": 0.09055596590042114, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 1086 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.49609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1829.0, "completions/mean_length": 1437.201171875, "completions/mean_terminated_length": 835.8720703125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.37108474865579927, "frac_reward_zero_std": 0.71875, "grad_norm": 22.999905614479072, "learning_rate": 8.136786177550373e-07, "loss": 0.0133, "num_tokens": 1059606427.0, "reward": 0.08203125, "reward_std": 0.10418535768985748, "rewards/accuracy_reward/mean": 0.08203125, "rewards/accuracy_reward/std": 0.2746807038784027, "step": 1087 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1902.0, "completions/mean_length": 1672.20703125, "completions/mean_terminated_length": 902.7261962890625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.3714261329691901, "frac_reward_zero_std": 0.78125, "grad_norm": 6.359560024218893, "learning_rate": 8.132439981092364e-07, "loss": 0.0113, "num_tokens": 1060537765.0, "reward": 0.033203125, "reward_std": 0.07493096590042114, "rewards/accuracy_reward/mean": 0.033203125, "rewards/accuracy_reward/std": 0.17934183776378632, "step": 1088 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.5859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1974.0, "completions/mean_length": 1543.43359375, "completions/mean_terminated_length": 829.7028198242188, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.37176751728258084, "frac_reward_zero_std": 0.78125, "grad_norm": 6.374684591557838, "learning_rate": 8.128090048367283e-07, "loss": 0.0082, "num_tokens": 1061404211.0, "reward": 0.072265625, "reward_std": 0.07885697484016418, "rewards/accuracy_reward/mean": 0.072265625, "rewards/accuracy_reward/std": 0.2591804563999176, "step": 1089 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.58984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2041.0, "completions/mean_length": 1576.244140625, "completions/mean_terminated_length": 897.8143310546875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.37210890159597165, "frac_reward_zero_std": 0.6875, "grad_norm": 4.398562849048429, "learning_rate": 8.123736385549063e-07, "loss": 0.0319, "num_tokens": 1062285552.0, "reward": 0.078125, "reward_std": 0.12488233298063278, "rewards/accuracy_reward/mean": 0.078125, "rewards/accuracy_reward/std": 0.26863065361976624, "step": 1090 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2048.0, "completions/mean_length": 1670.970703125, "completions/mean_terminated_length": 925.6802368164062, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.37245028590936247, "frac_reward_zero_std": 0.78125, "grad_norm": 8.052033287637672, "learning_rate": 8.119378998816932e-07, "loss": 0.0187, "num_tokens": 1063222513.0, "reward": 0.048828125, "reward_std": 0.08406735956668854, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 1091 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.0, "completions/mean_length": 1693.908203125, "completions/mean_terminated_length": 789.201416015625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.3727916702227533, "frac_reward_zero_std": 0.875, "grad_norm": 8.729028892685118, "learning_rate": 8.115017894355401e-07, "loss": 0.0122, "num_tokens": 1064173650.0, "reward": 0.056640625, "reward_std": 0.05833513289690018, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 1092 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.650390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1979.0, "completions/mean_length": 1629.107421875, "completions/mean_terminated_length": 849.8267822265625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.37313305453614404, "frac_reward_zero_std": 0.78125, "grad_norm": 2.8854574359168157, "learning_rate": 8.110653078354264e-07, "loss": 0.0123, "num_tokens": 1065081177.0, "reward": 0.041015625, "reward_std": 0.07085913419723511, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 1093 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.70703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1959.0, "completions/mean_length": 1704.26171875, "completions/mean_terminated_length": 874.7066650390625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.37347443884953485, "frac_reward_zero_std": 0.84375, "grad_norm": 5.242001679256016, "learning_rate": 8.106284557008577e-07, "loss": 0.0033, "num_tokens": 1066031247.0, "reward": 0.072265625, "reward_std": 0.064970001578331, "rewards/accuracy_reward/mean": 0.072265625, "rewards/accuracy_reward/std": 0.2591804563999176, "step": 1094 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1932.0, "completions/mean_length": 1841.87109375, "completions/mean_terminated_length": 791.5952758789062, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.37381582316292566, "frac_reward_zero_std": 0.8125, "grad_norm": 4.040588492249451, "learning_rate": 8.101912336518656e-07, "loss": 0.0277, "num_tokens": 1067052205.0, "reward": 0.0234375, "reward_std": 0.06068410724401474, "rewards/accuracy_reward/mean": 0.0234375, "rewards/accuracy_reward/std": 0.15143637359142303, "step": 1095 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.712890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2033.0, "completions/mean_length": 1712.34765625, "completions/mean_terminated_length": 878.9251708984375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.3741572074763165, "frac_reward_zero_std": 0.8125, "grad_norm": 16.9268085170734, "learning_rate": 8.097536423090072e-07, "loss": 0.0057, "num_tokens": 1068006719.0, "reward": 0.041015625, "reward_std": 0.059305962175130844, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 1096 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.740234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1973.0, "completions/mean_length": 1705.68359375, "completions/mean_terminated_length": 730.2105712890625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.37449859178970724, "frac_reward_zero_std": 0.78125, "grad_norm": 15.510482943250402, "learning_rate": 8.09315682293363e-07, "loss": 0.0086, "num_tokens": 1068947501.0, "reward": 0.05859375, "reward_std": 0.0921671986579895, "rewards/accuracy_reward/mean": 0.05859375, "rewards/accuracy_reward/std": 0.23509246110916138, "step": 1097 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.822265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1909.0, "completions/mean_length": 1839.494140625, "completions/mean_terminated_length": 874.8681640625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.37483997610309805, "frac_reward_zero_std": 0.875, "grad_norm": 18.31251210113687, "learning_rate": 8.088773542265372e-07, "loss": 0.0029, "num_tokens": 1069966170.0, "reward": 0.029296875, "reward_std": 0.04672222584486008, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 1098 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.84375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1763.0, "completions/mean_length": 1848.9453125, "completions/mean_terminated_length": 774.0499877929688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.37518136041648886, "frac_reward_zero_std": 0.90625, "grad_norm": 4.000336071300132, "learning_rate": 8.084386587306566e-07, "loss": 0.004, "num_tokens": 1070994686.0, "reward": 0.02734375, "reward_std": 0.02960042469203472, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 1099 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.80078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2044.0, "completions/mean_length": 1823.26171875, "completions/mean_terminated_length": 920.421630859375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.3755227447298797, "frac_reward_zero_std": 0.8125, "grad_norm": 6.038236609404691, "learning_rate": 8.079995964283688e-07, "loss": 0.0159, "num_tokens": 1072005508.0, "reward": 0.0234375, "reward_std": 0.05880707502365112, "rewards/accuracy_reward/mean": 0.0234375, "rewards/accuracy_reward/std": 0.15143637359142303, "step": 1100 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.771484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1962.0, "completions/mean_length": 1767.470703125, "completions/mean_terminated_length": 820.3846435546875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.37586412904327043, "frac_reward_zero_std": 0.78125, "grad_norm": 5.828429751976278, "learning_rate": 8.075601679428427e-07, "loss": 0.031, "num_tokens": 1072985013.0, "reward": 0.029296875, "reward_std": 0.07085912674665451, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17417415976524353, "step": 1101 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2026.0, "completions/mean_length": 1836.1953125, "completions/mean_terminated_length": 869.2608642578125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.37620551335666125, "frac_reward_zero_std": 0.875, "grad_norm": 14.382093211937953, "learning_rate": 8.071203738977667e-07, "loss": 0.0205, "num_tokens": 1074005241.0, "reward": 0.021484375, "reward_std": 0.04792051762342453, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 1102 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.849609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1997.0, "completions/mean_length": 1866.09765625, "completions/mean_terminated_length": 838.467529296875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.37654689767005206, "frac_reward_zero_std": 0.84375, "grad_norm": 4.101733005141735, "learning_rate": 8.066802149173479e-07, "loss": 0.008, "num_tokens": 1075033451.0, "reward": 0.029296875, "reward_std": 0.05683041363954544, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 1103 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.814453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.0, "completions/mean_length": 1840.23046875, "completions/mean_terminated_length": 928.2316284179688, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.3768882819834429, "frac_reward_zero_std": 0.84375, "grad_norm": 6.716512237858535, "learning_rate": 8.062396916263112e-07, "loss": 0.0136, "num_tokens": 1076055233.0, "reward": 0.025390625, "reward_std": 0.05452118441462517, "rewards/accuracy_reward/mean": 0.025390625, "rewards/accuracy_reward/std": 0.15746226906776428, "step": 1104 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2020.0, "completions/mean_length": 1859.087890625, "completions/mean_terminated_length": 775.3289794921875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.3772296662968337, "frac_reward_zero_std": 0.90625, "grad_norm": 4.115055927360768, "learning_rate": 8.057988046498993e-07, "loss": 0.0134, "num_tokens": 1077078606.0, "reward": 0.009765625, "reward_std": 0.029160313308238983, "rewards/accuracy_reward/mean": 0.010080644860863686, "rewards/accuracy_reward/std": 0.0999959334731102, "step": 1105 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.80078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1948.0, "completions/mean_length": 1781.9765625, "completions/mean_terminated_length": 712.6666870117188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.37757105061022445, "frac_reward_zero_std": 0.90625, "grad_norm": 7.372664871952672, "learning_rate": 8.053575546138706e-07, "loss": 0.0007, "num_tokens": 1078061810.0, "reward": 0.005859375, "reward_std": 0.0234375, "rewards/accuracy_reward/mean": 0.005859375, "rewards/accuracy_reward/std": 0.07639661431312561, "step": 1106 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.78515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2005.0, "completions/mean_length": 1809.986328125, "completions/mean_terminated_length": 940.5363159179688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.37791243492361526, "frac_reward_zero_std": 0.90625, "grad_norm": 22.947873390003068, "learning_rate": 8.049159421444986e-07, "loss": 0.0058, "num_tokens": 1079076699.0, "reward": 0.029296875, "reward_std": 0.04177113622426987, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 1107 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.775390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1875.0, "completions/mean_length": 1721.556640625, "completions/mean_terminated_length": 594.6173706054688, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3782538192370061, "frac_reward_zero_std": 0.78125, "grad_norm": 7.1075845706332785, "learning_rate": 8.044739678685713e-07, "loss": 0.0162, "num_tokens": 1080033288.0, "reward": 0.046875, "reward_std": 0.07961063086986542, "rewards/accuracy_reward/mean": 0.046875, "rewards/accuracy_reward/std": 0.21157780289649963, "step": 1108 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1946.0, "completions/mean_length": 1788.685546875, "completions/mean_terminated_length": 903.4396362304688, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.3785952035503969, "frac_reward_zero_std": 0.625, "grad_norm": 8.756116931166353, "learning_rate": 8.040316324133907e-07, "loss": 0.0297, "num_tokens": 1081024999.0, "reward": 0.0859375, "reward_std": 0.13534268736839294, "rewards/accuracy_reward/mean": 0.0859375, "rewards/accuracy_reward/std": 0.28054583072662354, "step": 1109 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.81640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1989.0, "completions/mean_length": 1818.587890625, "completions/mean_terminated_length": 798.4361572265625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.37893658786378764, "frac_reward_zero_std": 0.8125, "grad_norm": 3.859190621081395, "learning_rate": 8.035889364067709e-07, "loss": -0.0043, "num_tokens": 1082036228.0, "reward": 0.033203125, "reward_std": 0.06661957502365112, "rewards/accuracy_reward/mean": 0.033203125, "rewards/accuracy_reward/std": 0.17934183776378632, "step": 1110 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.78125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1836.0, "completions/mean_length": 1774.75, "completions/mean_terminated_length": 798.857177734375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.37927797217717846, "frac_reward_zero_std": 0.78125, "grad_norm": 8.301070333254545, "learning_rate": 8.031458804770379e-07, "loss": 0.012, "num_tokens": 1083022692.0, "reward": 0.044921875, "reward_std": 0.08259069174528122, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 1111 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.75390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1973.0, "completions/mean_length": 1704.130859375, "completions/mean_terminated_length": 650.6904907226562, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.37961935649056927, "frac_reward_zero_std": 0.75, "grad_norm": 4.613399055731976, "learning_rate": 8.027024652530285e-07, "loss": 0.0069, "num_tokens": 1083971479.0, "reward": 0.056640625, "reward_std": 0.0830162987112999, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 1112 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1952.0, "completions/mean_length": 1849.041015625, "completions/mean_terminated_length": 835.297607421875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.3799607408039601, "frac_reward_zero_std": 0.78125, "grad_norm": 2.520319595351757, "learning_rate": 8.022586913640896e-07, "loss": 0.0265, "num_tokens": 1084995180.0, "reward": 0.025390625, "reward_std": 0.06536141037940979, "rewards/accuracy_reward/mean": 0.025390625, "rewards/accuracy_reward/std": 0.15746226906776428, "step": 1113 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.77734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2043.0, "completions/mean_length": 1809.962890625, "completions/mean_terminated_length": 978.9210815429688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.38030212511735084, "frac_reward_zero_std": 0.78125, "grad_norm": 7.777566763334661, "learning_rate": 8.018145594400772e-07, "loss": 0.0257, "num_tokens": 1085989129.0, "reward": 0.072265625, "reward_std": 0.08699656277894974, "rewards/accuracy_reward/mean": 0.072265625, "rewards/accuracy_reward/std": 0.2591804563999176, "step": 1114 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.728515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2037.0, "completions/mean_length": 1691.48046875, "completions/mean_terminated_length": 734.7769775390625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.38064350943074166, "frac_reward_zero_std": 0.84375, "grad_norm": 16.463706565937134, "learning_rate": 8.01370070111355e-07, "loss": 0.0163, "num_tokens": 1086931791.0, "reward": 0.029296875, "reward_std": 0.055033616721630096, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 1115 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.712890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1924.0, "completions/mean_length": 1646.90234375, "completions/mean_terminated_length": 650.9795532226562, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.38098489374413247, "frac_reward_zero_std": 0.875, "grad_norm": 1.6077925663347319, "learning_rate": 8.009252240087947e-07, "loss": 0.0086, "num_tokens": 1087849421.0, "reward": 0.021484375, "reward_std": 0.04230976849794388, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 1116 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.744140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2025.0, "completions/mean_length": 1739.02734375, "completions/mean_terminated_length": 840.801513671875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3813262780575233, "frac_reward_zero_std": 0.8125, "grad_norm": 36.13097715464304, "learning_rate": 8.004800217637736e-07, "loss": 0.0193, "num_tokens": 1088827403.0, "reward": 0.052734375, "reward_std": 0.06755761802196503, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 1117 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.75, "completions/max_length": 2048.0, "completions/max_terminated_length": 1995.0, "completions/mean_length": 1716.001953125, "completions/mean_terminated_length": 720.0078125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.38166766237091404, "frac_reward_zero_std": 0.78125, "grad_norm": 12.039312582270002, "learning_rate": 8.000344640081752e-07, "loss": 0.0111, "num_tokens": 1089779164.0, "reward": 0.064453125, "reward_std": 0.08654290437698364, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 1118 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.75, "completions/max_length": 2048.0, "completions/max_terminated_length": 1977.0, "completions/mean_length": 1743.798828125, "completions/mean_terminated_length": 831.1953125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.38200904668430485, "frac_reward_zero_std": 0.8125, "grad_norm": 9.104278326846828, "learning_rate": 7.99588551374387e-07, "loss": 0.0074, "num_tokens": 1090746357.0, "reward": 0.03125, "reward_std": 0.06469620764255524, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 1119 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.708984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2026.0, "completions/mean_length": 1657.306640625, "completions/mean_terminated_length": 705.4832153320312, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.38235043099769567, "frac_reward_zero_std": 0.5625, "grad_norm": 17.49240402776864, "learning_rate": 7.991422844953004e-07, "loss": 0.0412, "num_tokens": 1091671378.0, "reward": 0.10546875, "reward_std": 0.1713758409023285, "rewards/accuracy_reward/mean": 0.10546875, "rewards/accuracy_reward/std": 0.3074568510055542, "step": 1120 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.720703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2037.0, "completions/mean_length": 1724.23828125, "completions/mean_terminated_length": 888.7971801757812, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.3826918153110865, "frac_reward_zero_std": 0.78125, "grad_norm": 6.098260673817675, "learning_rate": 7.986956640043096e-07, "loss": 0.0292, "num_tokens": 1092635964.0, "reward": 0.029296875, "reward_std": 0.07328042387962341, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 1121 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.73046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2018.0, "completions/mean_length": 1707.16796875, "completions/mean_terminated_length": 783.4638061523438, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.38303319962447724, "frac_reward_zero_std": 0.90625, "grad_norm": 4.022968640160011, "learning_rate": 7.98248690535311e-07, "loss": 0.0079, "num_tokens": 1093581250.0, "reward": 0.005859375, "reward_std": 0.0234375, "rewards/accuracy_reward/mean": 0.006048386916518211, "rewards/accuracy_reward/std": 0.07761410623788834, "step": 1122 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2035.0, "completions/mean_length": 1673.734375, "completions/mean_terminated_length": 879.5609130859375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.38337458393786805, "frac_reward_zero_std": 0.75, "grad_norm": 10.632064039184149, "learning_rate": 7.978013647227015e-07, "loss": 0.0185, "num_tokens": 1094517114.0, "reward": 0.044921875, "reward_std": 0.0930916965007782, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 1123 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.708984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1906.0, "completions/mean_length": 1657.404296875, "completions/mean_terminated_length": 705.8187866210938, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.38371596825125887, "frac_reward_zero_std": 0.84375, "grad_norm": 5.0890692605384205, "learning_rate": 7.973536872013783e-07, "loss": 0.0059, "num_tokens": 1095435881.0, "reward": 0.04296875, "reward_std": 0.057157501578330994, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 1124 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1996.0, "completions/mean_length": 1699.830078125, "completions/mean_terminated_length": 961.0304565429688, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.3840573525646497, "frac_reward_zero_std": 0.8125, "grad_norm": 15.258839064005763, "learning_rate": 7.969056586067376e-07, "loss": -0.0113, "num_tokens": 1096383778.0, "reward": 0.02734375, "reward_std": 0.05806133896112442, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 1125 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1929.0, "completions/mean_length": 1649.353515625, "completions/mean_terminated_length": 772.3312377929688, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.38439873687804044, "frac_reward_zero_std": 0.75, "grad_norm": 11.698117812687881, "learning_rate": 7.964572795746741e-07, "loss": 0.0118, "num_tokens": 1097297655.0, "reward": 0.037109375, "reward_std": 0.08560486882925034, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 1126 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.740234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1990.0, "completions/mean_length": 1703.35546875, "completions/mean_terminated_length": 721.2481689453125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.38474012119143125, "frac_reward_zero_std": 0.78125, "grad_norm": 12.440638392115558, "learning_rate": 7.960085507415802e-07, "loss": 0.0209, "num_tokens": 1098251661.0, "reward": 0.0859375, "reward_std": 0.08482664078474045, "rewards/accuracy_reward/mean": 0.0859375, "rewards/accuracy_reward/std": 0.28054583072662354, "step": 1127 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2041.0, "completions/mean_length": 1707.841796875, "completions/mean_terminated_length": 767.4044189453125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.38508150550482206, "frac_reward_zero_std": 0.78125, "grad_norm": 17.644220379694776, "learning_rate": 7.955594727443439e-07, "loss": 0.0175, "num_tokens": 1099202876.0, "reward": 0.046875, "reward_std": 0.08419393002986908, "rewards/accuracy_reward/mean": 0.046875, "rewards/accuracy_reward/std": 0.21157780289649963, "step": 1128 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.74609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2010.0, "completions/mean_length": 1707.509765625, "completions/mean_terminated_length": 706.9923095703125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.3854228898182129, "frac_reward_zero_std": 0.8125, "grad_norm": 13.601851376862413, "learning_rate": 7.951100462203494e-07, "loss": -0.0001, "num_tokens": 1100146257.0, "reward": 0.03515625, "reward_std": 0.05880707502365112, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 1129 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.68359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1930.0, "completions/mean_length": 1623.9140625, "completions/mean_terminated_length": 707.6790161132812, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.38576427413160363, "frac_reward_zero_std": 0.75, "grad_norm": 7.154470532031593, "learning_rate": 7.946602718074756e-07, "loss": 0.0221, "num_tokens": 1101065013.0, "reward": 0.056640625, "reward_std": 0.08686304092407227, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 1130 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.73046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2045.0, "completions/mean_length": 1723.4609375, "completions/mean_terminated_length": 844.3115844726562, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.38610565844499445, "frac_reward_zero_std": 0.8125, "grad_norm": 21.538006287101485, "learning_rate": 7.94210150144095e-07, "loss": 0.0034, "num_tokens": 1102023649.0, "reward": 0.064453125, "reward_std": 0.08074785768985748, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 1131 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.666015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1961.0, "completions/mean_length": 1616.080078125, "completions/mean_terminated_length": 754.76611328125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.38644704275838526, "frac_reward_zero_std": 0.84375, "grad_norm": 1.008773652447448, "learning_rate": 7.937596818690729e-07, "loss": 0.0063, "num_tokens": 1102927226.0, "reward": 0.025390625, "reward_std": 0.05024883896112442, "rewards/accuracy_reward/mean": 0.025390625, "rewards/accuracy_reward/std": 0.15746226906776428, "step": 1132 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.751953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2045.0, "completions/mean_length": 1745.6171875, "completions/mean_terminated_length": 828.9448852539062, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3867884270717761, "frac_reward_zero_std": 0.875, "grad_norm": 3.3852533880674023, "learning_rate": 7.933088676217667e-07, "loss": 0.0056, "num_tokens": 1103907910.0, "reward": 0.013671875, "reward_std": 0.03889618441462517, "rewards/accuracy_reward/mean": 0.01411290280520916, "rewards/accuracy_reward/std": 0.11807556450366974, "step": 1133 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.732421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1937.0, "completions/mean_length": 1696.43359375, "completions/mean_terminated_length": 734.1167602539062, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.38712981138516683, "frac_reward_zero_std": 0.78125, "grad_norm": 10.653070292176036, "learning_rate": 7.928577080420247e-07, "loss": 0.001, "num_tokens": 1104855892.0, "reward": 0.0703125, "reward_std": 0.08235205709934235, "rewards/accuracy_reward/mean": 0.0703125, "rewards/accuracy_reward/std": 0.25592297315597534, "step": 1134 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.77734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1986.0, "completions/mean_length": 1772.447265625, "completions/mean_terminated_length": 810.4298095703125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.38747119569855765, "frac_reward_zero_std": 0.78125, "grad_norm": 22.529968173272234, "learning_rate": 7.924062037701853e-07, "loss": 0.0035, "num_tokens": 1105848041.0, "reward": 0.044921875, "reward_std": 0.08928529918193817, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 1135 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.697265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2028.0, "completions/mean_length": 1685.462890625, "completions/mean_terminated_length": 850.4580688476562, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.38781258001194846, "frac_reward_zero_std": 0.71875, "grad_norm": 7.81079984928334, "learning_rate": 7.919543554470763e-07, "loss": 0.0122, "num_tokens": 1106794550.0, "reward": 0.03125, "reward_std": 0.08549975603818893, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 1136 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.66015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1948.0, "completions/mean_length": 1610.267578125, "completions/mean_terminated_length": 759.9597778320312, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.3881539643253393, "frac_reward_zero_std": 0.71875, "grad_norm": 15.103790470330274, "learning_rate": 7.91502163714014e-07, "loss": 0.0129, "num_tokens": 1107698831.0, "reward": 0.0859375, "reward_std": 0.09739763289690018, "rewards/accuracy_reward/mean": 0.0859375, "rewards/accuracy_reward/std": 0.28054583072662354, "step": 1137 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.583984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2030.0, "completions/mean_length": 1492.013671875, "completions/mean_terminated_length": 711.5446166992188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.38849534863873003, "frac_reward_zero_std": 0.8125, "grad_norm": 3.6574998971869004, "learning_rate": 7.910496292128015e-07, "loss": 0.0177, "num_tokens": 1108538678.0, "reward": 0.07421875, "reward_std": 0.07619608938694, "rewards/accuracy_reward/mean": 0.07421875, "rewards/accuracy_reward/std": 0.2623828947544098, "step": 1138 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.693359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2043.0, "completions/mean_length": 1679.5703125, "completions/mean_terminated_length": 846.496826171875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.38883673295212084, "frac_reward_zero_std": 0.8125, "grad_norm": 17.651934307100472, "learning_rate": 7.905967525857291e-07, "loss": -0.0023, "num_tokens": 1109469162.0, "reward": 0.03125, "reward_std": 0.0637841522693634, "rewards/accuracy_reward/mean": 0.032258063554763794, "rewards/accuracy_reward/std": 0.17686307430267334, "step": 1139 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.75, "completions/max_length": 2048.0, "completions/max_terminated_length": 1943.0, "completions/mean_length": 1748.384765625, "completions/mean_terminated_length": 849.5390625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.38917811726551166, "frac_reward_zero_std": 0.875, "grad_norm": 5.882594443243401, "learning_rate": 7.901435344755721e-07, "loss": 0.0087, "num_tokens": 1110442271.0, "reward": 0.015625, "reward_std": 0.040274329483509064, "rewards/accuracy_reward/mean": 0.015625, "rewards/accuracy_reward/std": 0.12414088100194931, "step": 1140 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.65234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2007.0, "completions/mean_length": 1598.85546875, "completions/mean_terminated_length": 756.0786743164062, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.38951950157890247, "frac_reward_zero_std": 0.84375, "grad_norm": 8.576852939157652, "learning_rate": 7.896899755255906e-07, "loss": 0.0012, "num_tokens": 1111336773.0, "reward": 0.044921875, "reward_std": 0.05193261057138443, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 1141 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.59765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2043.0, "completions/mean_length": 1552.900390625, "completions/mean_terminated_length": 817.461181640625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.38986088589229323, "frac_reward_zero_std": 0.6875, "grad_norm": 26.516191919548543, "learning_rate": 7.892360763795291e-07, "loss": 0.015, "num_tokens": 1112203474.0, "reward": 0.095703125, "reward_std": 0.10746511071920395, "rewards/accuracy_reward/mean": 0.095703125, "rewards/accuracy_reward/std": 0.2944713830947876, "step": 1142 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.626953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1994.0, "completions/mean_length": 1582.880859375, "completions/mean_terminated_length": 801.481689453125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.39020227020568404, "frac_reward_zero_std": 0.5625, "grad_norm": 10.336358793659462, "learning_rate": 7.887818376816136e-07, "loss": 0.0025, "num_tokens": 1113096757.0, "reward": 0.12109375, "reward_std": 0.16460810601711273, "rewards/accuracy_reward/mean": 0.12109375, "rewards/accuracy_reward/std": 0.3265552520751953, "step": 1143 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.62890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2025.0, "completions/mean_length": 1548.65625, "completions/mean_terminated_length": 702.678955078125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.39054365451907486, "frac_reward_zero_std": 0.75, "grad_norm": 16.005698554614572, "learning_rate": 7.883272600765535e-07, "loss": 0.0077, "num_tokens": 1113979045.0, "reward": 0.041015625, "reward_std": 0.08070705831050873, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 1144 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.642578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2029.0, "completions/mean_length": 1610.380859375, "completions/mean_terminated_length": 823.6229248046875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.39088503883246567, "frac_reward_zero_std": 0.875, "grad_norm": 1.206436976535436, "learning_rate": 7.878723442095384e-07, "loss": 0.0011, "num_tokens": 1114881880.0, "reward": 0.029296875, "reward_std": 0.04125870764255524, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 1145 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.658203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.0, "completions/mean_length": 1622.7109375, "completions/mean_terminated_length": 804.0171508789062, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.39122642314585643, "frac_reward_zero_std": 0.71875, "grad_norm": 8.636780994474929, "learning_rate": 7.87417090726238e-07, "loss": -0.005, "num_tokens": 1115795124.0, "reward": 0.083984375, "reward_std": 0.10530310869216919, "rewards/accuracy_reward/mean": 0.083984375, "rewards/accuracy_reward/std": 0.2776356339454651, "step": 1146 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1603.263671875, "completions/mean_terminated_length": 836.8031616210938, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.39156780745924724, "frac_reward_zero_std": 0.8125, "grad_norm": 7.3195402867773485, "learning_rate": 7.869615002728016e-07, "loss": 0.0106, "num_tokens": 1116688171.0, "reward": 0.05859375, "reward_std": 0.06079617515206337, "rewards/accuracy_reward/mean": 0.05859375, "rewards/accuracy_reward/std": 0.23509246110916138, "step": 1147 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.62890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1874.0, "completions/mean_length": 1587.91015625, "completions/mean_terminated_length": 808.47900390625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.39190919177263805, "frac_reward_zero_std": 0.71875, "grad_norm": 6.080562278427425, "learning_rate": 7.865055734958566e-07, "loss": 0.015, "num_tokens": 1117581021.0, "reward": 0.041015625, "reward_std": 0.09479551017284393, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 1148 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.630859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1945.0, "completions/mean_length": 1554.697265625, "completions/mean_terminated_length": 711.6455078125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.39225057608602887, "frac_reward_zero_std": 0.59375, "grad_norm": 10.614844030817949, "learning_rate": 7.860493110425073e-07, "loss": 0.0413, "num_tokens": 1118454418.0, "reward": 0.130859375, "reward_std": 0.14998990297317505, "rewards/accuracy_reward/mean": 0.130859375, "rewards/accuracy_reward/std": 0.33757632970809937, "step": 1149 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.611328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2026.0, "completions/mean_length": 1547.953125, "completions/mean_terminated_length": 761.4472045898438, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.3925919603994196, "frac_reward_zero_std": 0.65625, "grad_norm": 41.61745241700161, "learning_rate": 7.855927135603348e-07, "loss": 0.0022, "num_tokens": 1119319866.0, "reward": 0.0625, "reward_std": 0.11923094093799591, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.2422981858253479, "step": 1150 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.552734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1944.0, "completions/mean_length": 1529.962890625, "completions/mean_terminated_length": 889.7685546875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.39293334471281044, "frac_reward_zero_std": 0.5625, "grad_norm": 28.764453255424353, "learning_rate": 7.851357816973962e-07, "loss": 0.0198, "num_tokens": 1120174695.0, "reward": 0.111328125, "reward_std": 0.16661281883716583, "rewards/accuracy_reward/mean": 0.11491935700178146, "rewards/accuracy_reward/std": 0.3192465901374817, "step": 1151 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1965.0, "completions/mean_length": 1523.212890625, "completions/mean_terminated_length": 730.8872680664062, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.39327472902620125, "frac_reward_zero_std": 0.75, "grad_norm": 9.035769089590765, "learning_rate": 7.846785161022223e-07, "loss": 0.0244, "num_tokens": 1121038036.0, "reward": 0.044921875, "reward_std": 0.09126891195774078, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 1152 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.646484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1935.0, "completions/mean_length": 1624.607421875, "completions/mean_terminated_length": 850.3370361328125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.39361611333959207, "frac_reward_zero_std": 0.8125, "grad_norm": 18.997149227823183, "learning_rate": 7.842209174238181e-07, "loss": -0.0066, "num_tokens": 1121952331.0, "reward": 0.041015625, "reward_std": 0.06354551017284393, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 1153 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.63671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2032.0, "completions/mean_length": 1607.54296875, "completions/mean_terminated_length": 835.5591430664062, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.3939574976529828, "frac_reward_zero_std": 0.84375, "grad_norm": 10.485292639226472, "learning_rate": 7.837629863116612e-07, "loss": 0.0056, "num_tokens": 1122843681.0, "reward": 0.041015625, "reward_std": 0.06018522381782532, "rewards/accuracy_reward/mean": 0.04233871027827263, "rewards/accuracy_reward/std": 0.2015640139579773, "step": 1154 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1983.0, "completions/mean_length": 1661.201171875, "completions/mean_terminated_length": 778.5064086914062, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.39429888196637364, "frac_reward_zero_std": 0.8125, "grad_norm": 10.591285810002844, "learning_rate": 7.833047234157012e-07, "loss": 0.0203, "num_tokens": 1123770280.0, "reward": 0.060546875, "reward_std": 0.06733252108097076, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 1155 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1963.0, "completions/mean_length": 1641.947265625, "completions/mean_terminated_length": 748.6312866210938, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.39464026627976445, "frac_reward_zero_std": 0.875, "grad_norm": 3.008939650255724, "learning_rate": 7.828461293863581e-07, "loss": 0.0079, "num_tokens": 1124691533.0, "reward": 0.017578125, "reward_std": 0.04125870764255524, "rewards/accuracy_reward/mean": 0.017578125, "rewards/accuracy_reward/std": 0.13154059648513794, "step": 1156 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.666015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1943.0, "completions/mean_length": 1631.20703125, "completions/mean_terminated_length": 800.3099365234375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.39498165059315526, "frac_reward_zero_std": 0.71875, "grad_norm": 15.803212358734505, "learning_rate": 7.823872048745223e-07, "loss": 0.0136, "num_tokens": 1125612999.0, "reward": 0.060546875, "reward_std": 0.10740634053945541, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 1157 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.669921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1991.0, "completions/mean_length": 1627.06640625, "completions/mean_terminated_length": 772.7455444335938, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.395323034906546, "frac_reward_zero_std": 0.84375, "grad_norm": 10.240800015547556, "learning_rate": 7.819279505315538e-07, "loss": 0.0217, "num_tokens": 1126521497.0, "reward": 0.0390625, "reward_std": 0.06355906277894974, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 1158 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2018.0, "completions/mean_length": 1677.330078125, "completions/mean_terminated_length": 692.4071655273438, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.39566441921993684, "frac_reward_zero_std": 0.8125, "grad_norm": 1.2449334635798142, "learning_rate": 7.814683670092795e-07, "loss": 0.0109, "num_tokens": 1127453522.0, "reward": 0.029296875, "reward_std": 0.059984706342220306, "rewards/accuracy_reward/mean": 0.030241934582591057, "rewards/accuracy_reward/std": 0.1714252382516861, "step": 1159 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.728515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1970.0, "completions/mean_length": 1713.470703125, "completions/mean_terminated_length": 816.1582641601562, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.39600580353332765, "frac_reward_zero_std": 0.8125, "grad_norm": 9.58854134045238, "learning_rate": 7.810084549599944e-07, "loss": 0.0131, "num_tokens": 1128414339.0, "reward": 0.025390625, "reward_std": 0.06260748207569122, "rewards/accuracy_reward/mean": 0.026209676638245583, "rewards/accuracy_reward/std": 0.1599196344614029, "step": 1160 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.76171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2045.0, "completions/mean_length": 1724.216796875, "completions/mean_terminated_length": 689.172119140625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.39634718784671846, "frac_reward_zero_std": 0.84375, "grad_norm": 6.310055790491502, "learning_rate": 7.805482150364598e-07, "loss": 0.0092, "num_tokens": 1129369154.0, "reward": 0.052734375, "reward_std": 0.0659080371260643, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 1161 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.70703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1951.0, "completions/mean_length": 1642.552734375, "completions/mean_terminated_length": 664.0733642578125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.3966885721601092, "frac_reward_zero_std": 0.75, "grad_norm": 21.13245130994666, "learning_rate": 7.800876478919014e-07, "loss": 0.0135, "num_tokens": 1130292237.0, "reward": 0.083984375, "reward_std": 0.10387972742319107, "rewards/accuracy_reward/mean": 0.083984375, "rewards/accuracy_reward/std": 0.2776356339454651, "step": 1162 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.705078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2005.0, "completions/mean_length": 1691.845703125, "completions/mean_terminated_length": 840.3775024414062, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.39702995647350003, "frac_reward_zero_std": 0.71875, "grad_norm": 11.272049032269972, "learning_rate": 7.796267541800106e-07, "loss": 0.0169, "num_tokens": 1131233982.0, "reward": 0.044921875, "reward_std": 0.09644509106874466, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 1163 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.705078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1946.0, "completions/mean_length": 1690.845703125, "completions/mean_terminated_length": 836.9867553710938, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.39737134078689085, "frac_reward_zero_std": 0.8125, "grad_norm": 18.326459236882656, "learning_rate": 7.791655345549416e-07, "loss": 0.0052, "num_tokens": 1132177183.0, "reward": 0.046875, "reward_std": 0.07041902095079422, "rewards/accuracy_reward/mean": 0.046875, "rewards/accuracy_reward/std": 0.21157780289649963, "step": 1164 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.693359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1884.0, "completions/mean_length": 1666.48046875, "completions/mean_terminated_length": 803.8088989257812, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.39771272510028166, "frac_reward_zero_std": 0.875, "grad_norm": 3.804194544750429, "learning_rate": 7.78703989671311e-07, "loss": 0.0034, "num_tokens": 1133099317.0, "reward": 0.037109375, "reward_std": 0.05574656277894974, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 1165 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1940.0, "completions/mean_length": 1733.892578125, "completions/mean_terminated_length": 751.040283203125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.3980541094136724, "frac_reward_zero_std": 0.78125, "grad_norm": 11.97469355275299, "learning_rate": 7.782421201841978e-07, "loss": 0.0197, "num_tokens": 1134062958.0, "reward": 0.033203125, "reward_std": 0.07712717354297638, "rewards/accuracy_reward/mean": 0.033203125, "rewards/accuracy_reward/std": 0.17934183776378632, "step": 1166 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1971.0, "completions/mean_length": 1727.140625, "completions/mean_terminated_length": 803.45458984375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.39839549372706323, "frac_reward_zero_std": 0.8125, "grad_norm": 7.930897395036539, "learning_rate": 7.77779926749141e-07, "loss": -0.0013, "num_tokens": 1135025222.0, "reward": 0.060546875, "reward_std": 0.08444078266620636, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 1167 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.669921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1839.0, "completions/mean_length": 1643.603515625, "completions/mean_terminated_length": 822.84619140625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.39873687804045405, "frac_reward_zero_std": 0.6875, "grad_norm": 19.743556297672896, "learning_rate": 7.773174100221398e-07, "loss": 0.0043, "num_tokens": 1135941467.0, "reward": 0.06640625, "reward_std": 0.12108887732028961, "rewards/accuracy_reward/mean": 0.06640625, "rewards/accuracy_reward/std": 0.2492343932390213, "step": 1168 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2033.0, "completions/mean_length": 1630.64453125, "completions/mean_terminated_length": 745.0365600585938, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.39907826235384486, "frac_reward_zero_std": 0.71875, "grad_norm": 7.1051099731647325, "learning_rate": 7.768545706596519e-07, "loss": 0.0124, "num_tokens": 1136850405.0, "reward": 0.1015625, "reward_std": 0.11063909530639648, "rewards/accuracy_reward/mean": 0.1015625, "rewards/accuracy_reward/std": 0.30236753821372986, "step": 1169 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.724609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2020.0, "completions/mean_length": 1690.173828125, "completions/mean_terminated_length": 748.6595458984375, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.3994196466672356, "frac_reward_zero_std": 0.8125, "grad_norm": 4.318624409156783, "learning_rate": 7.763914093185932e-07, "loss": 0.01, "num_tokens": 1137790430.0, "reward": 0.033203125, "reward_std": 0.06849660724401474, "rewards/accuracy_reward/mean": 0.03427419438958168, "rewards/accuracy_reward/std": 0.18211629986763, "step": 1170 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.736328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1982.0, "completions/mean_length": 1718.1640625, "completions/mean_terminated_length": 797.066650390625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.39976103098062643, "frac_reward_zero_std": 0.6875, "grad_norm": 16.22759870386561, "learning_rate": 7.759279266563365e-07, "loss": 0.024, "num_tokens": 1138750658.0, "reward": 0.044921875, "reward_std": 0.09877577424049377, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 1171 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.751953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1941.0, "completions/mean_length": 1696.5703125, "completions/mean_terminated_length": 631.2125854492188, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.40010241529401724, "frac_reward_zero_std": 0.84375, "grad_norm": 3.1614684579522003, "learning_rate": 7.754641233307109e-07, "loss": 0.008, "num_tokens": 1139702598.0, "reward": 0.017578125, "reward_std": 0.04957009106874466, "rewards/accuracy_reward/mean": 0.018145160749554634, "rewards/accuracy_reward/std": 0.1336110234260559, "step": 1172 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2016.0, "completions/mean_length": 1658.375, "completions/mean_terminated_length": 888.18603515625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.40044379960740806, "frac_reward_zero_std": 0.8125, "grad_norm": 9.293519585649774, "learning_rate": 7.75e-07, "loss": 0.0182, "num_tokens": 1140632998.0, "reward": 0.037109375, "reward_std": 0.07174387574195862, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 1173 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.642578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1899.0, "completions/mean_length": 1617.62109375, "completions/mean_terminated_length": 843.8797607421875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.4007851839207988, "frac_reward_zero_std": 0.65625, "grad_norm": 20.616121201423113, "learning_rate": 7.745355573229422e-07, "loss": 0.0203, "num_tokens": 1141533828.0, "reward": 0.0625, "reward_std": 0.11075212061405182, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.2422981858253479, "step": 1174 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.677734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1998.0, "completions/mean_length": 1619.31640625, "completions/mean_terminated_length": 717.7817993164062, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.40112656823418963, "frac_reward_zero_std": 0.78125, "grad_norm": 3.249735378161522, "learning_rate": 7.740707959587289e-07, "loss": 0.0222, "num_tokens": 1142437238.0, "reward": 0.03125, "reward_std": 0.07426576316356659, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 1175 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.654296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2031.0, "completions/mean_length": 1614.1484375, "completions/mean_terminated_length": 793.0169677734375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.40146795254758044, "frac_reward_zero_std": 0.71875, "grad_norm": 7.441774705816416, "learning_rate": 7.736057165670038e-07, "loss": 0.0173, "num_tokens": 1143337890.0, "reward": 0.0625, "reward_std": 0.09749060869216919, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.2422981858253479, "step": 1176 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.68359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1891.0, "completions/mean_length": 1663.08984375, "completions/mean_terminated_length": 831.4938354492188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.40180933686097126, "frac_reward_zero_std": 0.71875, "grad_norm": 15.209668395812065, "learning_rate": 7.731403198078622e-07, "loss": 0.0284, "num_tokens": 1144263808.0, "reward": 0.060546875, "reward_std": 0.10194281488656998, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 1177 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1866.0, "completions/mean_length": 1722.638671875, "completions/mean_terminated_length": 785.992431640625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.402150721174362, "frac_reward_zero_std": 0.8125, "grad_norm": 7.136172237557808, "learning_rate": 7.726746063418493e-07, "loss": 0.013, "num_tokens": 1145225975.0, "reward": 0.07421875, "reward_std": 0.06409768760204315, "rewards/accuracy_reward/mean": 0.07421875, "rewards/accuracy_reward/std": 0.2623828947544098, "step": 1178 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.701171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1980.0, "completions/mean_length": 1645.662109375, "completions/mean_terminated_length": 701.6143798828125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.4024921054877528, "frac_reward_zero_std": 0.875, "grad_norm": 3.537834559232754, "learning_rate": 7.722085768299608e-07, "loss": 0.0198, "num_tokens": 1146144058.0, "reward": 0.013671875, "reward_std": 0.03889618441462517, "rewards/accuracy_reward/mean": 0.013671875, "rewards/accuracy_reward/std": 0.1162383034825325, "step": 1179 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.775390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2018.0, "completions/mean_length": 1777.16796875, "completions/mean_terminated_length": 842.2086791992188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.40283348980114364, "frac_reward_zero_std": 0.6875, "grad_norm": 14.960035798261194, "learning_rate": 7.717422319336398e-07, "loss": 0.0185, "num_tokens": 1147132240.0, "reward": 0.046875, "reward_std": 0.09825544059276581, "rewards/accuracy_reward/mean": 0.046875, "rewards/accuracy_reward/std": 0.21157780289649963, "step": 1180 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.724609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1988.0, "completions/mean_length": 1726.03515625, "completions/mean_terminated_length": 878.87939453125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.40317487411453445, "frac_reward_zero_std": 0.875, "grad_norm": 5.938458838567134, "learning_rate": 7.712755723147777e-07, "loss": 0.0015, "num_tokens": 1148085234.0, "reward": 0.0390625, "reward_std": 0.042695626616477966, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 1181 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1989.0, "completions/mean_length": 1684.74609375, "completions/mean_terminated_length": 774.123291015625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.4035162584279252, "frac_reward_zero_std": 0.71875, "grad_norm": 9.398862745165575, "learning_rate": 7.708085986357127e-07, "loss": 0.0047, "num_tokens": 1149024624.0, "reward": 0.037109375, "reward_std": 0.09055596590042114, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 1182 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.76953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2009.0, "completions/mean_length": 1759.21484375, "completions/mean_terminated_length": 794.9661254882812, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.403857642741316, "frac_reward_zero_std": 0.75, "grad_norm": 1.8903023179523921, "learning_rate": 7.703413115592282e-07, "loss": 0.0093, "num_tokens": 1150004734.0, "reward": 0.04296875, "reward_std": 0.0801548957824707, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 1183 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.748046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1875.0, "completions/mean_length": 1735.970703125, "completions/mean_terminated_length": 809.5581665039062, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.40419902705470684, "frac_reward_zero_std": 0.6875, "grad_norm": 14.877807766225922, "learning_rate": 7.69873711748553e-07, "loss": 0.0089, "num_tokens": 1150972911.0, "reward": 0.05859375, "reward_std": 0.1154462918639183, "rewards/accuracy_reward/mean": 0.05859375, "rewards/accuracy_reward/std": 0.23509246110916138, "step": 1184 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.80078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1984.0, "completions/mean_length": 1794.025390625, "completions/mean_terminated_length": 773.7059326171875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.40454041136809765, "frac_reward_zero_std": 0.8125, "grad_norm": 32.604778904381504, "learning_rate": 7.694057998673597e-07, "loss": 0.0031, "num_tokens": 1151975052.0, "reward": 0.0390625, "reward_std": 0.072782501578331, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 1185 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.798828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1990.0, "completions/mean_length": 1805.548828125, "completions/mean_terminated_length": 842.8058471679688, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.4048817956814884, "frac_reward_zero_std": 0.875, "grad_norm": 4.141016035438558, "learning_rate": 7.68937576579763e-07, "loss": 0.0168, "num_tokens": 1152978485.0, "reward": 0.029296875, "reward_std": 0.05193261057138443, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 1186 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1709.0, "completions/mean_length": 1694.248046875, "completions/mean_terminated_length": 675.8712158203125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.4052231799948792, "frac_reward_zero_std": 0.71875, "grad_norm": 13.324399305135795, "learning_rate": 7.684690425503208e-07, "loss": 0.0106, "num_tokens": 1153925444.0, "reward": 0.08984375, "reward_std": 0.1293414831161499, "rewards/accuracy_reward/mean": 0.08984375, "rewards/accuracy_reward/std": 0.2862374484539032, "step": 1187 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.82421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1966.0, "completions/mean_length": 1824.5625, "completions/mean_terminated_length": 776.888916015625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.40556456430827004, "frac_reward_zero_std": 0.8125, "grad_norm": 18.18535227366748, "learning_rate": 7.680001984440312e-07, "loss": 0.0125, "num_tokens": 1154932276.0, "reward": 0.025390625, "reward_std": 0.06354551762342453, "rewards/accuracy_reward/mean": 0.02708333358168602, "rewards/accuracy_reward/std": 0.16249564290046692, "step": 1188 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1817.0, "completions/mean_length": 1858.6015625, "completions/mean_terminated_length": 621.941162109375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.40590594862166085, "frac_reward_zero_std": 0.8125, "grad_norm": 8.118619357088836, "learning_rate": 7.675310449263326e-07, "loss": 0.0221, "num_tokens": 1155963352.0, "reward": 0.046875, "reward_std": 0.06794346868991852, "rewards/accuracy_reward/mean": 0.046875, "rewards/accuracy_reward/std": 0.21157780289649963, "step": 1189 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.87109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2036.0, "completions/mean_length": 1896.361328125, "completions/mean_terminated_length": 871.6515502929688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.4062473329350516, "frac_reward_zero_std": 0.84375, "grad_norm": 12.603861080594282, "learning_rate": 7.670615826631027e-07, "loss": 0.0186, "num_tokens": 1157003121.0, "reward": 0.05078125, "reward_std": 0.054249756038188934, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 1190 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.83984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2004.0, "completions/mean_length": 1841.23828125, "completions/mean_terminated_length": 757.0, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.4065887172484424, "frac_reward_zero_std": 0.8125, "grad_norm": 10.359905530122068, "learning_rate": 7.665918123206572e-07, "loss": 0.0193, "num_tokens": 1158016075.0, "reward": 0.03515625, "reward_std": 0.06233368441462517, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 1191 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1837.0, "completions/mean_length": 1881.796875, "completions/mean_terminated_length": 629.7333374023438, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.40693010156183324, "frac_reward_zero_std": 0.78125, "grad_norm": 11.872453086380027, "learning_rate": 7.661217345657495e-07, "loss": 0.0091, "num_tokens": 1159053523.0, "reward": 0.025390625, "reward_std": 0.06661957502365112, "rewards/accuracy_reward/mean": 0.025390625, "rewards/accuracy_reward/std": 0.15746226906776428, "step": 1192 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1919.0, "completions/mean_length": 1769.171875, "completions/mean_terminated_length": 620.3999633789062, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.40727148587522405, "frac_reward_zero_std": 0.84375, "grad_norm": 7.550368853248177, "learning_rate": 7.656513500655688e-07, "loss": 0.0157, "num_tokens": 1160033131.0, "reward": 0.03515625, "reward_std": 0.06233368441462517, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 1193 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.87890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2039.0, "completions/mean_length": 1891.80078125, "completions/mean_terminated_length": 758.0967407226562, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.4076128701886148, "frac_reward_zero_std": 0.84375, "grad_norm": 13.799740948957593, "learning_rate": 7.651806594877397e-07, "loss": 0.0097, "num_tokens": 1161074389.0, "reward": 0.04296875, "reward_std": 0.07014063745737076, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 1194 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.837890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2011.0, "completions/mean_length": 1838.001953125, "completions/mean_terminated_length": 752.59033203125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.4079542545020056, "frac_reward_zero_std": 0.78125, "grad_norm": 6.285437921261953, "learning_rate": 7.647096635003216e-07, "loss": 0.0225, "num_tokens": 1162098710.0, "reward": 0.021484375, "reward_std": 0.06519509106874466, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 1195 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.0, "completions/mean_length": 1877.72265625, "completions/mean_terminated_length": 765.9117431640625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.40829563881539643, "frac_reward_zero_std": 0.8125, "grad_norm": 11.30567212016571, "learning_rate": 7.642383627718072e-07, "loss": 0.0024, "num_tokens": 1163137304.0, "reward": 0.02734375, "reward_std": 0.06640692055225372, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 1196 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.81640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1845.0, "completions/mean_length": 1798.380859375, "completions/mean_terminated_length": 688.372314453125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.40863702312878725, "frac_reward_zero_std": 0.71875, "grad_norm": 17.882567851948014, "learning_rate": 7.637667579711215e-07, "loss": 0.0017, "num_tokens": 1164143275.0, "reward": 0.048828125, "reward_std": 0.09448197484016418, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 1197 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1957.0, "completions/mean_length": 1824.7734375, "completions/mean_terminated_length": 749.227294921875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.408978407442178, "frac_reward_zero_std": 0.84375, "grad_norm": 21.30616430554781, "learning_rate": 7.632948497676213e-07, "loss": -0.0036, "num_tokens": 1165153223.0, "reward": 0.04296875, "reward_std": 0.050948236137628555, "rewards/accuracy_reward/mean": 0.04435483738780022, "rewards/accuracy_reward/std": 0.2060900777578354, "step": 1198 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.865234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1735.0, "completions/mean_length": 1867.1796875, "completions/mean_terminated_length": 706.2608642578125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.4093197917555688, "frac_reward_zero_std": 0.75, "grad_norm": 10.651026897248558, "learning_rate": 7.62822638831094e-07, "loss": 0.0139, "num_tokens": 1166178675.0, "reward": 0.03515625, "reward_std": 0.08345640450716019, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 1199 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.826171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2007.0, "completions/mean_length": 1781.787109375, "completions/mean_terminated_length": 516.528076171875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.40966117606895963, "frac_reward_zero_std": 0.875, "grad_norm": 4.70211317942562, "learning_rate": 7.623501258317567e-07, "loss": 0.009, "num_tokens": 1167171110.0, "reward": 0.025390625, "reward_std": 0.04863205552101135, "rewards/accuracy_reward/mean": 0.025390625, "rewards/accuracy_reward/std": 0.15746226906776428, "step": 1200 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.841796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1875.0, "completions/mean_length": 1849.70703125, "completions/mean_terminated_length": 794.5925903320312, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.41000256038235044, "frac_reward_zero_std": 0.78125, "grad_norm": 20.545322889086556, "learning_rate": 7.618773114402554e-07, "loss": 0.0146, "num_tokens": 1168192336.0, "reward": 0.056640625, "reward_std": 0.08483455330133438, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 1201 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.896484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1974.0, "completions/mean_length": 1916.453125, "completions/mean_terminated_length": 777.2075805664062, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.41034394469574126, "frac_reward_zero_std": 0.90625, "grad_norm": 2.702355764053578, "learning_rate": 7.61404196327663e-07, "loss": 0.0067, "num_tokens": 1169249848.0, "reward": 0.009765625, "reward_std": 0.029160313308238983, "rewards/accuracy_reward/mean": 0.009765625, "rewards/accuracy_reward/std": 0.09843364357948303, "step": 1202 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.88671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1936.0, "completions/mean_length": 1914.017578125, "completions/mean_terminated_length": 865.2586059570312, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.410685329009132, "frac_reward_zero_std": 0.875, "grad_norm": 3.4252063974196933, "learning_rate": 7.609307811654804e-07, "loss": 0.0067, "num_tokens": 1170311105.0, "reward": 0.013671875, "reward_std": 0.03889618441462517, "rewards/accuracy_reward/mean": 0.013671875, "rewards/accuracy_reward/std": 0.1162383034825325, "step": 1203 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.85546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1704.0, "completions/mean_length": 1833.22265625, "completions/mean_terminated_length": 561.9729614257812, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.41102671332252283, "frac_reward_zero_std": 0.8125, "grad_norm": 13.684863938742346, "learning_rate": 7.604570666256339e-07, "loss": 0.0065, "num_tokens": 1171328355.0, "reward": 0.025390625, "reward_std": 0.06409074366092682, "rewards/accuracy_reward/mean": 0.025390625, "rewards/accuracy_reward/std": 0.15746226906776428, "step": 1204 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.798828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2034.0, "completions/mean_length": 1783.40625, "completions/mean_terminated_length": 732.7378540039062, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.41136809763591364, "frac_reward_zero_std": 0.75, "grad_norm": 37.810948958138965, "learning_rate": 7.599830533804741e-07, "loss": 0.0129, "num_tokens": 1172319763.0, "reward": 0.0703125, "reward_std": 0.08126020431518555, "rewards/accuracy_reward/mean": 0.07258064299821854, "rewards/accuracy_reward/std": 0.25970885157585144, "step": 1205 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.849609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2030.0, "completions/mean_length": 1840.693359375, "completions/mean_terminated_length": 669.5454711914062, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.41170948194930446, "frac_reward_zero_std": 0.78125, "grad_norm": 7.747585638411618, "learning_rate": 7.595087421027767e-07, "loss": 0.0052, "num_tokens": 1173338854.0, "reward": 0.021484375, "reward_std": 0.06519509106874466, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 1206 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.826171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1948.0, "completions/mean_length": 1808.953125, "completions/mean_terminated_length": 672.8090209960938, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.4120508662626952, "frac_reward_zero_std": 0.84375, "grad_norm": 8.741608849899395, "learning_rate": 7.590341334657395e-07, "loss": 0.0041, "num_tokens": 1174343806.0, "reward": 0.015625, "reward_std": 0.04670868441462517, "rewards/accuracy_reward/mean": 0.015625, "rewards/accuracy_reward/std": 0.12414088100194931, "step": 1207 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.818359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1785.0, "completions/mean_length": 1798.60546875, "completions/mean_terminated_length": 674.9892578125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.41239225057608603, "frac_reward_zero_std": 0.75, "grad_norm": 45.23357192343699, "learning_rate": 7.585592281429828e-07, "loss": 0.0115, "num_tokens": 1175344868.0, "reward": 0.080078125, "reward_std": 0.08830691128969193, "rewards/accuracy_reward/mean": 0.080078125, "rewards/accuracy_reward/std": 0.271679550409317, "step": 1208 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.826171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2022.0, "completions/mean_length": 1834.201171875, "completions/mean_terminated_length": 818.0562133789062, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.41273363488947684, "frac_reward_zero_std": 0.75, "grad_norm": 14.144580254330286, "learning_rate": 7.580840268085477e-07, "loss": 0.0082, "num_tokens": 1176360635.0, "reward": 0.046875, "reward_std": 0.08461953699588776, "rewards/accuracy_reward/mean": 0.046875, "rewards/accuracy_reward/std": 0.21157780289649963, "step": 1209 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.861328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1986.0, "completions/mean_length": 1869.185546875, "completions/mean_terminated_length": 758.5211181640625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.41307501920286765, "frac_reward_zero_std": 0.875, "grad_norm": 6.035850002135212, "learning_rate": 7.576085301368955e-07, "loss": 0.0021, "num_tokens": 1177387706.0, "reward": 0.01953125, "reward_std": 0.04318207502365112, "rewards/accuracy_reward/mean": 0.01953125, "rewards/accuracy_reward/std": 0.1385180652141571, "step": 1210 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.810546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1948.0, "completions/mean_length": 1800.435546875, "completions/mean_terminated_length": 741.7937622070312, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.4134164035162584, "frac_reward_zero_std": 0.875, "grad_norm": 27.616297749082623, "learning_rate": 7.571327388029071e-07, "loss": 0.0109, "num_tokens": 1178396137.0, "reward": 0.0546875, "reward_std": 0.05507335811853409, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 1211 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.865234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1960.0, "completions/mean_length": 1854.47265625, "completions/mean_terminated_length": 611.9710083007812, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.4137577878296492, "frac_reward_zero_std": 0.78125, "grad_norm": 15.147621101863834, "learning_rate": 7.566566534818809e-07, "loss": 0.0155, "num_tokens": 1179423995.0, "reward": 0.03515625, "reward_std": 0.0720367580652237, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 1212 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.841796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1960.0, "completions/mean_length": 1846.283203125, "completions/mean_terminated_length": 773.6049194335938, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.41409917214304004, "frac_reward_zero_std": 0.9375, "grad_norm": 8.862783353806183, "learning_rate": 7.561802748495332e-07, "loss": 0.0094, "num_tokens": 1180446972.0, "reward": 0.017578125, "reward_std": 0.02394993044435978, "rewards/accuracy_reward/mean": 0.017578125, "rewards/accuracy_reward/std": 0.13154059648513794, "step": 1213 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.861328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1891.0, "completions/mean_length": 1863.234375, "completions/mean_terminated_length": 715.6055908203125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.41444055645643085, "frac_reward_zero_std": 0.875, "grad_norm": 12.062186271184311, "learning_rate": 7.557036035819963e-07, "loss": 0.0064, "num_tokens": 1181476308.0, "reward": 0.017578125, "reward_std": 0.04461899772286415, "rewards/accuracy_reward/mean": 0.017578125, "rewards/accuracy_reward/std": 0.13154059648513794, "step": 1214 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2040.0, "completions/mean_length": 1867.73828125, "completions/mean_terminated_length": 766.138916015625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.4147819407698216, "frac_reward_zero_std": 0.8125, "grad_norm": 4.686802188307055, "learning_rate": 7.552266403558176e-07, "loss": 0.0223, "num_tokens": 1182504462.0, "reward": 0.03125, "reward_std": 0.0628461092710495, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 1215 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.90625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1837.0, "completions/mean_length": 1901.052734375, "completions/mean_terminated_length": 480.5625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.4151233250832124, "frac_reward_zero_std": 0.84375, "grad_norm": 11.723084811874012, "learning_rate": 7.547493858479595e-07, "loss": 0.0134, "num_tokens": 1183556329.0, "reward": 0.021484375, "reward_std": 0.05287160724401474, "rewards/accuracy_reward/mean": 0.02217741869390011, "rewards/accuracy_reward/std": 0.14740891754627228, "step": 1216 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.83984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1961.0, "completions/mean_length": 1848.107421875, "completions/mean_terminated_length": 799.8901977539062, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.41546470939660324, "frac_reward_zero_std": 0.875, "grad_norm": 19.674664914925824, "learning_rate": 7.542718407357973e-07, "loss": 0.0059, "num_tokens": 1184576592.0, "reward": 0.033203125, "reward_std": 0.048086829483509064, "rewards/accuracy_reward/mean": 0.033203125, "rewards/accuracy_reward/std": 0.17934183776378632, "step": 1217 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.84375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1828.0, "completions/mean_length": 1822.568359375, "completions/mean_terminated_length": 605.2374877929688, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.41580609370999405, "frac_reward_zero_std": 0.75, "grad_norm": 16.289511590722295, "learning_rate": 7.537940056971192e-07, "loss": 0.0078, "num_tokens": 1185594819.0, "reward": 0.044921875, "reward_std": 0.07344770431518555, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 1218 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.845703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1948.0, "completions/mean_length": 1841.654296875, "completions/mean_terminated_length": 710.6708984375, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.4161474780233848, "frac_reward_zero_std": 0.84375, "grad_norm": 3.947440805747222, "learning_rate": 7.533158814101242e-07, "loss": 0.011, "num_tokens": 1186623634.0, "reward": 0.01171875, "reward_std": 0.04192390665411949, "rewards/accuracy_reward/mean": 0.012096773833036423, "rewards/accuracy_reward/std": 0.10942844301462173, "step": 1219 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.83203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1993.0, "completions/mean_length": 1836.693359375, "completions/mean_terminated_length": 789.9883422851562, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.4164888623367756, "frac_reward_zero_std": 0.84375, "grad_norm": 13.528352408750054, "learning_rate": 7.528374685534227e-07, "loss": 0.0024, "num_tokens": 1187642677.0, "reward": 0.01953125, "reward_std": 0.04907120764255524, "rewards/accuracy_reward/mean": 0.01953125, "rewards/accuracy_reward/std": 0.1385180652141571, "step": 1220 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2009.0, "completions/mean_length": 1737.8671875, "completions/mean_terminated_length": 724.7667236328125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.41683024665016644, "frac_reward_zero_std": 0.71875, "grad_norm": 23.163101537814786, "learning_rate": 7.52358767806034e-07, "loss": 0.0037, "num_tokens": 1188609505.0, "reward": 0.044921875, "reward_std": 0.08505964279174805, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 1221 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2005.0, "completions/mean_length": 1792.892578125, "completions/mean_terminated_length": 741.8499755859375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.41717163096355725, "frac_reward_zero_std": 0.8125, "grad_norm": 77.34105892107726, "learning_rate": 7.518797798473865e-07, "loss": 0.0164, "num_tokens": 1189602250.0, "reward": 0.056640625, "reward_std": 0.07763300091028214, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 1222 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.783203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1986.0, "completions/mean_length": 1770.810546875, "completions/mean_terminated_length": 769.909912109375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.417513015276948, "frac_reward_zero_std": 0.84375, "grad_norm": 55.796782244880035, "learning_rate": 7.514005053573156e-07, "loss": 0.0149, "num_tokens": 1190588585.0, "reward": 0.041015625, "reward_std": 0.06150216609239578, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 1223 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.794921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2033.0, "completions/mean_length": 1753.103515625, "completions/mean_terminated_length": 610.028564453125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.4178543995903388, "frac_reward_zero_std": 0.90625, "grad_norm": 22.08431494552066, "learning_rate": 7.509209450160639e-07, "loss": 0.0133, "num_tokens": 1191558974.0, "reward": 0.0234375, "reward_std": 0.040274329483509064, "rewards/accuracy_reward/mean": 0.0234375, "rewards/accuracy_reward/std": 0.15143637359142303, "step": 1224 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.740234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1975.0, "completions/mean_length": 1704.06640625, "completions/mean_terminated_length": 723.9849853515625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.41819578390372963, "frac_reward_zero_std": 0.75, "grad_norm": 5.392557113231142, "learning_rate": 7.5044109950428e-07, "loss": 0.0298, "num_tokens": 1192506496.0, "reward": 0.0390625, "reward_std": 0.08792105317115784, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 1225 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.74609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2010.0, "completions/mean_length": 1712.748046875, "completions/mean_terminated_length": 727.6231079101562, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.41853716821712045, "frac_reward_zero_std": 0.84375, "grad_norm": 16.80782432403552, "learning_rate": 7.499609695030163e-07, "loss": 0.0158, "num_tokens": 1193461759.0, "reward": 0.02734375, "reward_std": 0.0580955371260643, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 1226 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.732421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1997.0, "completions/mean_length": 1707.0234375, "completions/mean_terminated_length": 773.6934204101562, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.4188785525305112, "frac_reward_zero_std": 0.84375, "grad_norm": 7.149140957121258, "learning_rate": 7.494805556937299e-07, "loss": 0.019, "num_tokens": 1194413115.0, "reward": 0.0390625, "reward_std": 0.05864076316356659, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 1227 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.72265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2030.0, "completions/mean_length": 1695.64453125, "completions/mean_terminated_length": 777.5352172851562, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.419219936843902, "frac_reward_zero_std": 0.6875, "grad_norm": 4.5494958734166975, "learning_rate": 7.4899985875828e-07, "loss": 0.0328, "num_tokens": 1195363493.0, "reward": 0.0625, "reward_std": 0.1155911535024643, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.2422981858253479, "step": 1228 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2020.0, "completions/mean_length": 1670.998046875, "completions/mean_terminated_length": 778.0986938476562, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.41956132115729283, "frac_reward_zero_std": 0.6875, "grad_norm": 14.959410518513367, "learning_rate": 7.485188793789284e-07, "loss": 0.0277, "num_tokens": 1196298580.0, "reward": 0.076171875, "reward_std": 0.12836134433746338, "rewards/accuracy_reward/mean": 0.076171875, "rewards/accuracy_reward/std": 0.26553234457969666, "step": 1229 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.701171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1953.0, "completions/mean_length": 1670.828125, "completions/mean_terminated_length": 785.830078125, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.41990270547068365, "frac_reward_zero_std": 0.6875, "grad_norm": 11.215384823305884, "learning_rate": 7.480376182383371e-07, "loss": 0.0387, "num_tokens": 1197235148.0, "reward": 0.0390625, "reward_std": 0.103924959897995, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 1230 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2009.0, "completions/mean_length": 1716.232421875, "completions/mean_terminated_length": 884.5410766601562, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.4202440897840744, "frac_reward_zero_std": 0.6875, "grad_norm": 22.159552435547813, "learning_rate": 7.47556076019568e-07, "loss": 0.0341, "num_tokens": 1198189891.0, "reward": 0.099609375, "reward_std": 0.12160685658454895, "rewards/accuracy_reward/mean": 0.099609375, "rewards/accuracy_reward/std": 0.29977133870124817, "step": 1231 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.705078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1975.0, "completions/mean_length": 1648.89453125, "completions/mean_terminated_length": 694.74169921875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.4205854740974652, "frac_reward_zero_std": 0.71875, "grad_norm": 17.960560402574, "learning_rate": 7.470742534060827e-07, "loss": 0.0076, "num_tokens": 1199124349.0, "reward": 0.099609375, "reward_std": 0.11199785768985748, "rewards/accuracy_reward/mean": 0.10282257944345474, "rewards/accuracy_reward/std": 0.30403366684913635, "step": 1232 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2016.0, "completions/mean_length": 1719.783203125, "completions/mean_terminated_length": 647.6083374023438, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.42092685841085603, "frac_reward_zero_std": 0.96875, "grad_norm": 27.99823521103077, "learning_rate": 7.465921510817401e-07, "loss": -0.0038, "num_tokens": 1200081262.0, "reward": 0.005859375, "reward_std": 0.012597277760505676, "rewards/accuracy_reward/mean": 0.005859375, "rewards/accuracy_reward/std": 0.07639661431312561, "step": 1233 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1964.0, "completions/mean_length": 1695.138671875, "completions/mean_terminated_length": 827.2905883789062, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.42126824272424684, "frac_reward_zero_std": 0.71875, "grad_norm": 12.443622654828356, "learning_rate": 7.461097697307962e-07, "loss": 0.0346, "num_tokens": 1201022805.0, "reward": 0.052734375, "reward_std": 0.08890638500452042, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 1234 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.79296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2015.0, "completions/mean_length": 1779.7890625, "completions/mean_terminated_length": 752.4906005859375, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.4216096270376376, "frac_reward_zero_std": 0.8125, "grad_norm": 9.760850151810333, "learning_rate": 7.456271100379031e-07, "loss": 0.0191, "num_tokens": 1202018265.0, "reward": 0.033203125, "reward_std": 0.062347229570150375, "rewards/accuracy_reward/mean": 0.033203125, "rewards/accuracy_reward/std": 0.17934183776378632, "step": 1235 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.751953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1988.0, "completions/mean_length": 1735.333984375, "completions/mean_terminated_length": 787.93701171875, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.4219510113510284, "frac_reward_zero_std": 0.59375, "grad_norm": 26.717921857520043, "learning_rate": 7.451441726881082e-07, "loss": 0.0011, "num_tokens": 1203000644.0, "reward": 0.109375, "reward_std": 0.15267950296401978, "rewards/accuracy_reward/mean": 0.109375, "rewards/accuracy_reward/std": 0.31241437792778015, "step": 1236 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1898.0, "completions/mean_length": 1654.96484375, "completions/mean_terminated_length": 758.0384521484375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.42229239566441923, "frac_reward_zero_std": 0.75, "grad_norm": 8.933408814328024, "learning_rate": 7.446609583668522e-07, "loss": 0.0249, "num_tokens": 1203924962.0, "reward": 0.04296875, "reward_std": 0.09302500635385513, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 1237 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.654296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1807.0, "completions/mean_length": 1587.53515625, "completions/mean_terminated_length": 716.0338745117188, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.42263377997781004, "frac_reward_zero_std": 0.6875, "grad_norm": 16.089814264083646, "learning_rate": 7.441774677599699e-07, "loss": 0.0158, "num_tokens": 1204816356.0, "reward": 0.0625, "reward_std": 0.11647041887044907, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.2422981858253479, "step": 1238 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.724609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1956.0, "completions/mean_length": 1680.64453125, "completions/mean_terminated_length": 714.0567016601562, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.4229751642912008, "frac_reward_zero_std": 0.84375, "grad_norm": 9.241428258729558, "learning_rate": 7.436937015536876e-07, "loss": 0.01, "num_tokens": 1205752974.0, "reward": 0.04296875, "reward_std": 0.06640692055225372, "rewards/accuracy_reward/mean": 0.04435483738780022, "rewards/accuracy_reward/std": 0.2060900777578354, "step": 1239 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.69140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2033.0, "completions/mean_length": 1664.947265625, "completions/mean_terminated_length": 806.7152099609375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.4233165486045916, "frac_reward_zero_std": 0.875, "grad_norm": 18.900943510274026, "learning_rate": 7.432096604346231e-07, "loss": 0.0024, "num_tokens": 1206685235.0, "reward": 0.03125, "reward_std": 0.04847269132733345, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 1240 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.712890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1813.0, "completions/mean_length": 1680.84375, "completions/mean_terminated_length": 769.197265625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.4236579329179824, "frac_reward_zero_std": 0.71875, "grad_norm": 17.89006391086748, "learning_rate": 7.427253450897844e-07, "loss": 0.0217, "num_tokens": 1207622259.0, "reward": 0.064453125, "reward_std": 0.10678637027740479, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 1241 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.685546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2020.0, "completions/mean_length": 1678.734375, "completions/mean_terminated_length": 873.689453125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.42399931723137324, "frac_reward_zero_std": 0.6875, "grad_norm": 13.237277116594331, "learning_rate": 7.422407562065678e-07, "loss": 0.0136, "num_tokens": 1208569067.0, "reward": 0.06640625, "reward_std": 0.10260801017284393, "rewards/accuracy_reward/mean": 0.06854838877916336, "rewards/accuracy_reward/std": 0.25293970108032227, "step": 1242 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.728515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1838.0, "completions/mean_length": 1696.89453125, "completions/mean_terminated_length": 754.7194213867188, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.424340701544764, "frac_reward_zero_std": 0.78125, "grad_norm": 15.79416028076382, "learning_rate": 7.417558944727593e-07, "loss": 0.0072, "num_tokens": 1209514645.0, "reward": 0.025390625, "reward_std": 0.06849661469459534, "rewards/accuracy_reward/mean": 0.025390625, "rewards/accuracy_reward/std": 0.15746226906776428, "step": 1243 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1991.0, "completions/mean_length": 1733.1796875, "completions/mean_terminated_length": 826.8787841796875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.4246820858581548, "frac_reward_zero_std": 0.65625, "grad_norm": 10.403694749853384, "learning_rate": 7.412707605765313e-07, "loss": 0.0142, "num_tokens": 1210481121.0, "reward": 0.07421875, "reward_std": 0.13016510009765625, "rewards/accuracy_reward/mean": 0.07421875, "rewards/accuracy_reward/std": 0.2623828947544098, "step": 1244 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1979.0, "completions/mean_length": 1673.142578125, "completions/mean_terminated_length": 817.69873046875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.4250234701715456, "frac_reward_zero_std": 0.6875, "grad_norm": 19.931124308886883, "learning_rate": 7.407853552064425e-07, "loss": 0.0137, "num_tokens": 1211414586.0, "reward": 0.08984375, "reward_std": 0.12862853705883026, "rewards/accuracy_reward/mean": 0.08984375, "rewards/accuracy_reward/std": 0.2862374484539032, "step": 1245 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.755859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1987.0, "completions/mean_length": 1724.74609375, "completions/mean_terminated_length": 723.9520263671875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.42536485448493644, "frac_reward_zero_std": 0.9375, "grad_norm": 1.4162949704939196, "learning_rate": 7.40299679051437e-07, "loss": 0.0058, "num_tokens": 1212385672.0, "reward": 0.033203125, "reward_std": 0.026572702452540398, "rewards/accuracy_reward/mean": 0.033203125, "rewards/accuracy_reward/std": 0.17934183776378632, "step": 1246 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.74609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1960.0, "completions/mean_length": 1743.173828125, "completions/mean_terminated_length": 847.8692016601562, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.4257062387983272, "frac_reward_zero_std": 0.8125, "grad_norm": 8.906044583163013, "learning_rate": 7.398137328008435e-07, "loss": 0.0106, "num_tokens": 1213358033.0, "reward": 0.048828125, "reward_std": 0.06645326316356659, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 1247 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.759765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1726.0, "completions/mean_length": 1738.42578125, "completions/mean_terminated_length": 759.3658447265625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.426047623111718, "frac_reward_zero_std": 0.75, "grad_norm": 20.936963334428164, "learning_rate": 7.393275171443737e-07, "loss": 0.0019, "num_tokens": 1214330475.0, "reward": 0.078125, "reward_std": 0.08989076316356659, "rewards/accuracy_reward/mean": 0.078125, "rewards/accuracy_reward/std": 0.26863065361976624, "step": 1248 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.75390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1741.0, "completions/mean_length": 1724.72265625, "completions/mean_terminated_length": 734.3651123046875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.4263890074251088, "frac_reward_zero_std": 0.78125, "grad_norm": 10.97586093646319, "learning_rate": 7.388410327721218e-07, "loss": 0.0041, "num_tokens": 1215285149.0, "reward": 0.037109375, "reward_std": 0.08082245290279388, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 1249 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.69921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1924.0, "completions/mean_length": 1684.994140625, "completions/mean_terminated_length": 841.1233520507812, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.42673039173849964, "frac_reward_zero_std": 0.78125, "grad_norm": 1.0239765281408597, "learning_rate": 7.383542803745632e-07, "loss": 0.0205, "num_tokens": 1216220250.0, "reward": 0.048828125, "reward_std": 0.07481793314218521, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 1250 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1994.0, "completions/mean_length": 1739.615234375, "completions/mean_terminated_length": 774.6693115234375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.4270717760518904, "frac_reward_zero_std": 0.6875, "grad_norm": 16.13745280223482, "learning_rate": 7.378672606425542e-07, "loss": 0.0187, "num_tokens": 1217182597.0, "reward": 0.05078125, "reward_std": 0.10178203880786896, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 1251 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2029.0, "completions/mean_length": 1722.560546875, "completions/mean_terminated_length": 822.816162109375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.4274131603652812, "frac_reward_zero_std": 0.78125, "grad_norm": 4.791048449799167, "learning_rate": 7.373799742673301e-07, "loss": 0.0253, "num_tokens": 1218136180.0, "reward": 0.044921875, "reward_std": 0.08110743016004562, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 1252 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1991.0, "completions/mean_length": 1700.58203125, "completions/mean_terminated_length": 566.0750122070312, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.427754544678672, "frac_reward_zero_std": 0.84375, "grad_norm": 33.165501688459926, "learning_rate": 7.36892421940505e-07, "loss": 0.0133, "num_tokens": 1219088462.0, "reward": 0.021484375, "reward_std": 0.05287160724401474, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 1253 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.779296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1809.0, "completions/mean_length": 1767.888671875, "completions/mean_terminated_length": 778.822998046875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.42809592899206284, "frac_reward_zero_std": 0.71875, "grad_norm": 16.834294690629708, "learning_rate": 7.364046043540699e-07, "loss": -0.0089, "num_tokens": 1220072645.0, "reward": 0.044921875, "reward_std": 0.09380322694778442, "rewards/accuracy_reward/mean": 0.0463709682226181, "rewards/accuracy_reward/std": 0.21049949526786804, "step": 1254 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.76953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2029.0, "completions/mean_length": 1720.9765625, "completions/mean_terminated_length": 629.0508422851562, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.4284373133054536, "frac_reward_zero_std": 0.8125, "grad_norm": 8.686193658310177, "learning_rate": 7.359165222003927e-07, "loss": 0.0221, "num_tokens": 1221031993.0, "reward": 0.037109375, "reward_std": 0.06508206576108932, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 1255 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.76171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1529.0, "completions/mean_length": 1723.244140625, "completions/mean_terminated_length": 685.0901489257812, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.4287786976188444, "frac_reward_zero_std": 0.71875, "grad_norm": 19.95496810582337, "learning_rate": 7.354281761722168e-07, "loss": 0.026, "num_tokens": 1221988486.0, "reward": 0.068359375, "reward_std": 0.10282756388187408, "rewards/accuracy_reward/mean": 0.068359375, "rewards/accuracy_reward/std": 0.25260838866233826, "step": 1256 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.69921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1997.0, "completions/mean_length": 1651.451171875, "completions/mean_terminated_length": 729.6038818359375, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.4291200819322352, "frac_reward_zero_std": 0.78125, "grad_norm": 2.6827198132044954, "learning_rate": 7.349395669626601e-07, "loss": 0.0257, "num_tokens": 1222908349.0, "reward": 0.03515625, "reward_std": 0.07053203880786896, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 1257 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.732421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1921.0, "completions/mean_length": 1728.71484375, "completions/mean_terminated_length": 854.7590942382812, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.42946146624562603, "frac_reward_zero_std": 0.84375, "grad_norm": 14.063289619289382, "learning_rate": 7.344506952652141e-07, "loss": 0.0034, "num_tokens": 1223869643.0, "reward": 0.02734375, "reward_std": 0.0580955371260643, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 1258 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.794921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1791.0, "completions/mean_length": 1761.693359375, "completions/mean_terminated_length": 651.914306640625, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.4298028505590168, "frac_reward_zero_std": 0.78125, "grad_norm": 7.025525419403623, "learning_rate": 7.339615617737427e-07, "loss": 0.0194, "num_tokens": 1224847694.0, "reward": 0.037109375, "reward_std": 0.07355976104736328, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 1259 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2042.0, "completions/mean_length": 1653.447265625, "completions/mean_terminated_length": 816.2255859375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.4301442348724076, "frac_reward_zero_std": 0.75, "grad_norm": 16.153004133961364, "learning_rate": 7.334721671824814e-07, "loss": 0.0042, "num_tokens": 1225775683.0, "reward": 0.041015625, "reward_std": 0.08648413419723511, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 1260 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.716796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1973.0, "completions/mean_length": 1673.494140625, "completions/mean_terminated_length": 725.6068725585938, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.4304856191857984, "frac_reward_zero_std": 0.90625, "grad_norm": 6.497636929099315, "learning_rate": 7.329825121860363e-07, "loss": 0.013, "num_tokens": 1226706960.0, "reward": 0.025390625, "reward_std": 0.030584799125790596, "rewards/accuracy_reward/mean": 0.026209676638245583, "rewards/accuracy_reward/std": 0.1599196344614029, "step": 1261 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.751953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1930.0, "completions/mean_length": 1702.712890625, "completions/mean_terminated_length": 655.9763793945312, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.43082700349918923, "frac_reward_zero_std": 0.8125, "grad_norm": 9.476483911043664, "learning_rate": 7.32492597479383e-07, "loss": 0.0055, "num_tokens": 1227656477.0, "reward": 0.037109375, "reward_std": 0.06860867142677307, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 1262 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.701171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1986.0, "completions/mean_length": 1664.287109375, "completions/mean_terminated_length": 763.941162109375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.43116838781258, "frac_reward_zero_std": 0.8125, "grad_norm": 8.427171632744047, "learning_rate": 7.32002423757866e-07, "loss": -0.0113, "num_tokens": 1228581536.0, "reward": 0.05078125, "reward_std": 0.06888246536254883, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 1263 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.728515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2021.0, "completions/mean_length": 1708.8203125, "completions/mean_terminated_length": 798.6475219726562, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.4315097721259708, "frac_reward_zero_std": 0.96875, "grad_norm": 1.1673621389787057, "learning_rate": 7.315119917171972e-07, "loss": -0.003, "num_tokens": 1229543076.0, "reward": 0.001953125, "reward_std": 0.0078125, "rewards/accuracy_reward/mean": 0.001953125, "rewards/accuracy_reward/std": 0.04419417306780815, "step": 1264 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.755859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1888.0, "completions/mean_length": 1700.34765625, "completions/mean_terminated_length": 624.0160522460938, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.4318511564393616, "frac_reward_zero_std": 0.8125, "grad_norm": 5.252421314944644, "learning_rate": 7.310213020534549e-07, "loss": 0.0118, "num_tokens": 1230493686.0, "reward": 0.07421875, "reward_std": 0.08709507435560226, "rewards/accuracy_reward/mean": 0.07421875, "rewards/accuracy_reward/std": 0.2623828947544098, "step": 1265 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.740234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2029.0, "completions/mean_length": 1708.05078125, "completions/mean_terminated_length": 739.3233032226562, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.43219254075275243, "frac_reward_zero_std": 0.8125, "grad_norm": 7.853572477672882, "learning_rate": 7.30530355463084e-07, "loss": 0.0117, "num_tokens": 1231449872.0, "reward": 0.02734375, "reward_std": 0.06327171623706818, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 1266 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.76171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2006.0, "completions/mean_length": 1758.078125, "completions/mean_terminated_length": 831.2786254882812, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.4325339250661432, "frac_reward_zero_std": 0.75, "grad_norm": 11.497914350827873, "learning_rate": 7.300391526428928e-07, "loss": 0.0148, "num_tokens": 1232435912.0, "reward": 0.064453125, "reward_std": 0.08341006934642792, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 1267 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1835.0, "completions/mean_length": 1661.224609375, "completions/mean_terminated_length": 633.5071411132812, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.432875309379534, "frac_reward_zero_std": 0.75, "grad_norm": 11.65346283302957, "learning_rate": 7.295476942900539e-07, "loss": 0.0108, "num_tokens": 1233364491.0, "reward": 0.1171875, "reward_std": 0.11482647061347961, "rewards/accuracy_reward/mean": 0.1171875, "rewards/accuracy_reward/std": 0.32195815443992615, "step": 1268 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.75, "completions/max_length": 2048.0, "completions/max_terminated_length": 1861.0, "completions/mean_length": 1731.416015625, "completions/mean_terminated_length": 781.6640625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.4332166936929248, "frac_reward_zero_std": 0.71875, "grad_norm": 43.932433158645814, "learning_rate": 7.290559811021029e-07, "loss": 0.0187, "num_tokens": 1234323680.0, "reward": 0.08984375, "reward_std": 0.09908141195774078, "rewards/accuracy_reward/mean": 0.08984375, "rewards/accuracy_reward/std": 0.2862374484539032, "step": 1269 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.837890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2025.0, "completions/mean_length": 1838.9375, "completions/mean_terminated_length": 758.3613891601562, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.43355807800631563, "frac_reward_zero_std": 0.75, "grad_norm": 11.863612608984905, "learning_rate": 7.285640137769363e-07, "loss": 0.0032, "num_tokens": 1235336672.0, "reward": 0.068359375, "reward_std": 0.09912915527820587, "rewards/accuracy_reward/mean": 0.068359375, "rewards/accuracy_reward/std": 0.25260838866233826, "step": 1270 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.798828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2001.0, "completions/mean_length": 1738.66015625, "completions/mean_terminated_length": 510.3106994628906, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.4338994623197064, "frac_reward_zero_std": 0.8125, "grad_norm": 8.047452249912592, "learning_rate": 7.280717930128119e-07, "loss": 0.0192, "num_tokens": 1236304130.0, "reward": 0.03515625, "reward_std": 0.06409768760204315, "rewards/accuracy_reward/mean": 0.03629032149910927, "rewards/accuracy_reward/std": 0.1872003972530365, "step": 1271 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.72265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2033.0, "completions/mean_length": 1688.8984375, "completions/mean_terminated_length": 753.2112426757812, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.4342408466330972, "frac_reward_zero_std": 0.8125, "grad_norm": 6.659324568222564, "learning_rate": 7.275793195083474e-07, "loss": -0.0073, "num_tokens": 1237247678.0, "reward": 0.0390625, "reward_std": 0.06469620764255524, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 1272 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.802734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1904.0, "completions/mean_length": 1781.5390625, "completions/mean_terminated_length": 697.2277221679688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.434582230946488, "frac_reward_zero_std": 0.75, "grad_norm": 13.547418837313424, "learning_rate": 7.270865939625183e-07, "loss": 0.0221, "num_tokens": 1238231730.0, "reward": 0.041015625, "reward_std": 0.08463403582572937, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 1273 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.716796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2022.0, "completions/mean_length": 1683.37109375, "completions/mean_terminated_length": 760.4827270507812, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.4349236152598788, "frac_reward_zero_std": 0.71875, "grad_norm": 19.91744369390336, "learning_rate": 7.265936170746588e-07, "loss": 0.0158, "num_tokens": 1239165424.0, "reward": 0.072265625, "reward_std": 0.10982763767242432, "rewards/accuracy_reward/mean": 0.072265625, "rewards/accuracy_reward/std": 0.2591804563999176, "step": 1274 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.79296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2033.0, "completions/mean_length": 1791.04296875, "completions/mean_terminated_length": 806.8490600585938, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.4352649995732696, "frac_reward_zero_std": 0.84375, "grad_norm": 8.707594556465457, "learning_rate": 7.261003895444593e-07, "loss": 0.008, "num_tokens": 1240150374.0, "reward": 0.0390625, "reward_std": 0.06299237906932831, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 1275 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2028.0, "completions/mean_length": 1651.208984375, "completions/mean_terminated_length": 656.513671875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.4356063838866604, "frac_reward_zero_std": 0.78125, "grad_norm": 8.7644118097855, "learning_rate": 7.256069120719661e-07, "loss": 0.0136, "num_tokens": 1241071825.0, "reward": 0.033203125, "reward_std": 0.0715966522693634, "rewards/accuracy_reward/mean": 0.033203125, "rewards/accuracy_reward/std": 0.17934183776378632, "step": 1276 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.775390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1924.0, "completions/mean_length": 1749.248046875, "completions/mean_terminated_length": 717.904296875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.4359477682000512, "frac_reward_zero_std": 0.78125, "grad_norm": 55.99714848488093, "learning_rate": 7.2511318535758e-07, "loss": 0.009, "num_tokens": 1242036880.0, "reward": 0.03125, "reward_std": 0.0752037987112999, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 1277 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.767578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1884.0, "completions/mean_length": 1742.1484375, "completions/mean_terminated_length": 732.0672607421875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.436289152513442, "frac_reward_zero_std": 0.84375, "grad_norm": 31.14684390893561, "learning_rate": 7.246192101020559e-07, "loss": -0.0057, "num_tokens": 1243006636.0, "reward": 0.0234375, "reward_std": 0.054249756038188934, "rewards/accuracy_reward/mean": 0.0234375, "rewards/accuracy_reward/std": 0.15143637359142303, "step": 1278 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.791015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1823.0, "completions/mean_length": 1774.8046875, "completions/mean_terminated_length": 740.7476196289062, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.4366305368268328, "frac_reward_zero_std": 0.78125, "grad_norm": 3.803038812629198, "learning_rate": 7.241249870065014e-07, "loss": 0.0116, "num_tokens": 1243994904.0, "reward": 0.0390625, "reward_std": 0.07477159798145294, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 1279 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.775390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2048.0, "completions/mean_length": 1779.265625, "completions/mean_terminated_length": 851.5477905273438, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.4369719211402236, "frac_reward_zero_std": 0.90625, "grad_norm": 1.2772127652322518, "learning_rate": 7.236305167723758e-07, "loss": 0.0086, "num_tokens": 1244976624.0, "reward": 0.015625, "reward_std": 0.03125, "rewards/accuracy_reward/mean": 0.015625, "rewards/accuracy_reward/std": 0.12414088100194931, "step": 1280 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.724609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2035.0, "completions/mean_length": 1697.666015625, "completions/mean_terminated_length": 776.2269287109375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.4373133054536144, "frac_reward_zero_std": 0.71875, "grad_norm": 24.919131172778428, "learning_rate": 7.231358001014891e-07, "loss": 0.0054, "num_tokens": 1245933365.0, "reward": 0.046875, "reward_std": 0.10409364104270935, "rewards/accuracy_reward/mean": 0.046875, "rewards/accuracy_reward/std": 0.21157780289649963, "step": 1281 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.755859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1970.0, "completions/mean_length": 1702.552734375, "completions/mean_terminated_length": 633.0480346679688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.4376546897670052, "frac_reward_zero_std": 0.90625, "grad_norm": 2.613020452400665, "learning_rate": 7.22640837696001e-07, "loss": 0.0115, "num_tokens": 1246879088.0, "reward": 0.0390625, "reward_std": 0.036034777760505676, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 1282 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.759765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2036.0, "completions/mean_length": 1734.173828125, "completions/mean_terminated_length": 741.6666259765625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.437996074080396, "frac_reward_zero_std": 0.90625, "grad_norm": 5.3903618508876585, "learning_rate": 7.221456302584202e-07, "loss": -0.0086, "num_tokens": 1247841753.0, "reward": 0.02734375, "reward_std": 0.02960042469203472, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 1283 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.775390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2007.0, "completions/mean_length": 1743.3984375, "completions/mean_terminated_length": 692.3043212890625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.4383374583937868, "frac_reward_zero_std": 0.75, "grad_norm": 11.52315698998046, "learning_rate": 7.216501784916032e-07, "loss": 0.0037, "num_tokens": 1248824757.0, "reward": 0.056640625, "reward_std": 0.09582037478685379, "rewards/accuracy_reward/mean": 0.058467742055654526, "rewards/accuracy_reward/std": 0.23486268520355225, "step": 1284 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.783203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1857.0, "completions/mean_length": 1774.2578125, "completions/mean_terminated_length": 785.3333740234375, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.4386788427071776, "frac_reward_zero_std": 0.8125, "grad_norm": 7.615453911609256, "learning_rate": 7.211544830987533e-07, "loss": 0.0264, "num_tokens": 1249804857.0, "reward": 0.080078125, "reward_std": 0.06900903582572937, "rewards/accuracy_reward/mean": 0.08266129344701767, "rewards/accuracy_reward/std": 0.2756475806236267, "step": 1285 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.744140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2026.0, "completions/mean_length": 1718.25390625, "completions/mean_terminated_length": 759.2213745117188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.4390202270205684, "frac_reward_zero_std": 0.90625, "grad_norm": 2.6303740504701167, "learning_rate": 7.206585447834188e-07, "loss": 0.015, "num_tokens": 1250763819.0, "reward": 0.015625, "reward_std": 0.03344620764255524, "rewards/accuracy_reward/mean": 0.015625, "rewards/accuracy_reward/std": 0.12414088100194931, "step": 1286 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.74609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1979.0, "completions/mean_length": 1716.93359375, "completions/mean_terminated_length": 744.107666015625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.4393616113339592, "frac_reward_zero_std": 0.8125, "grad_norm": 27.32345565839729, "learning_rate": 7.201623642494943e-07, "loss": 0.0142, "num_tokens": 1251718889.0, "reward": 0.03125, "reward_std": 0.06464291363954544, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 1287 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.68359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1887.0, "completions/mean_length": 1624.458984375, "completions/mean_terminated_length": 709.4012451171875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.43970299564735, "frac_reward_zero_std": 0.8125, "grad_norm": 10.467390542335078, "learning_rate": 7.196659422012166e-07, "loss": 0.0139, "num_tokens": 1252629892.0, "reward": 0.041015625, "reward_std": 0.06949453055858612, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 1288 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.755859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1799.0, "completions/mean_length": 1715.859375, "completions/mean_terminated_length": 687.5520629882812, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.4400443799607408, "frac_reward_zero_std": 0.96875, "grad_norm": 4.223577776104194, "learning_rate": 7.191692793431663e-07, "loss": -0.0037, "num_tokens": 1253584268.0, "reward": 0.0078125, "reward_std": 0.013975424692034721, "rewards/accuracy_reward/mean": 0.0078125, "rewards/accuracy_reward/std": 0.08812850713729858, "step": 1289 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.76171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1901.0, "completions/mean_length": 1726.337890625, "completions/mean_terminated_length": 698.07373046875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.4403857642741316, "frac_reward_zero_std": 0.78125, "grad_norm": 18.847524940092285, "learning_rate": 7.186723763802654e-07, "loss": -0.0003, "num_tokens": 1254543929.0, "reward": 0.03515625, "reward_std": 0.07317391037940979, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 1290 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1936.0, "completions/mean_length": 1777.169921875, "completions/mean_terminated_length": 764.5925903320312, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.4407271485875224, "frac_reward_zero_std": 0.78125, "grad_norm": 18.98281924602311, "learning_rate": 7.181752340177769e-07, "loss": 0.0236, "num_tokens": 1255540240.0, "reward": 0.037109375, "reward_std": 0.07669497281312943, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 1291 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.716796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1986.0, "completions/mean_length": 1676.091796875, "completions/mean_terminated_length": 734.779296875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.4410685329009132, "frac_reward_zero_std": 0.6875, "grad_norm": 25.14971518233946, "learning_rate": 7.176778529613035e-07, "loss": 0.0015, "num_tokens": 1256475679.0, "reward": 0.060546875, "reward_std": 0.11916647106409073, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 1292 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.751953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1929.0, "completions/mean_length": 1717.27734375, "completions/mean_terminated_length": 714.6929321289062, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.441409917214304, "frac_reward_zero_std": 0.78125, "grad_norm": 3.058746303821399, "learning_rate": 7.171802339167864e-07, "loss": 0.0171, "num_tokens": 1257431677.0, "reward": 0.03125, "reward_std": 0.07300759106874466, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 1293 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.810546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1823.0, "completions/mean_length": 1787.72265625, "completions/mean_terminated_length": 674.1649169921875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.4417513015276948, "frac_reward_zero_std": 0.84375, "grad_norm": 8.208561661090513, "learning_rate": 7.16682377590505e-07, "loss": 0.0212, "num_tokens": 1258421727.0, "reward": 0.0390625, "reward_std": 0.05507335811853409, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 1294 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1972.0, "completions/mean_length": 1705.705078125, "completions/mean_terminated_length": 720.3106079101562, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.4420926858410856, "frac_reward_zero_std": 0.78125, "grad_norm": 11.184418638972135, "learning_rate": 7.161842846890751e-07, "loss": 0.0257, "num_tokens": 1259371464.0, "reward": 0.046875, "reward_std": 0.06657323241233826, "rewards/accuracy_reward/mean": 0.046875, "rewards/accuracy_reward/std": 0.21157780289649963, "step": 1295 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.759765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1916.0, "completions/mean_length": 1738.138671875, "completions/mean_terminated_length": 758.5284423828125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.4424340701544764, "frac_reward_zero_std": 0.875, "grad_norm": 12.205306428292676, "learning_rate": 7.156859559194488e-07, "loss": 0.0112, "num_tokens": 1260339487.0, "reward": 0.037109375, "reward_std": 0.049345001578330994, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 1296 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.76953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1991.0, "completions/mean_length": 1735.060546875, "completions/mean_terminated_length": 690.1610107421875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.4427754544678672, "frac_reward_zero_std": 0.8125, "grad_norm": 16.701905559492975, "learning_rate": 7.151873919889122e-07, "loss": 0.0259, "num_tokens": 1261303566.0, "reward": 0.037109375, "reward_std": 0.07085912674665451, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 1297 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.783203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2003.0, "completions/mean_length": 1759.666015625, "completions/mean_terminated_length": 718.0270385742188, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.443116838781258, "frac_reward_zero_std": 0.8125, "grad_norm": 8.70065445300811, "learning_rate": 7.146885936050861e-07, "loss": 0.0185, "num_tokens": 1262282035.0, "reward": 0.052734375, "reward_std": 0.07586899399757385, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 1298 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.732421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2030.0, "completions/mean_length": 1690.265625, "completions/mean_terminated_length": 711.065673828125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.4434582230946488, "frac_reward_zero_std": 0.8125, "grad_norm": 7.899263490880784, "learning_rate": 7.141895614759232e-07, "loss": 0.0042, "num_tokens": 1263219835.0, "reward": 0.0390625, "reward_std": 0.07093241065740585, "rewards/accuracy_reward/mean": 0.04032257944345474, "rewards/accuracy_reward/std": 0.19691328704357147, "step": 1299 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1965.0, "completions/mean_length": 1700.80078125, "completions/mean_terminated_length": 740.8970336914062, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.4437996074080396, "frac_reward_zero_std": 0.75, "grad_norm": 11.908078900546084, "learning_rate": 7.136902963097085e-07, "loss": 0.0285, "num_tokens": 1264169509.0, "reward": 0.03515625, "reward_std": 0.08587770164012909, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 1300 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.755859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2001.0, "completions/mean_length": 1739.671875, "completions/mean_terminated_length": 785.0880126953125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.4441409917214304, "frac_reward_zero_std": 0.75, "grad_norm": 25.492125440100207, "learning_rate": 7.131907988150575e-07, "loss": 0.0083, "num_tokens": 1265146541.0, "reward": 0.046875, "reward_std": 0.09319227933883667, "rewards/accuracy_reward/mean": 0.046875, "rewards/accuracy_reward/std": 0.21157780289649963, "step": 1301 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.724609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1992.0, "completions/mean_length": 1703.83984375, "completions/mean_terminated_length": 798.28369140625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.4444823760348212, "frac_reward_zero_std": 0.71875, "grad_norm": 48.39820377490189, "learning_rate": 7.126910697009156e-07, "loss": 0.0238, "num_tokens": 1266100603.0, "reward": 0.0546875, "reward_std": 0.09859496355056763, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 1302 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.697265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2009.0, "completions/mean_length": 1646.345703125, "completions/mean_terminated_length": 721.2451782226562, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.444823760348212, "frac_reward_zero_std": 0.84375, "grad_norm": 2.6046642307956733, "learning_rate": 7.121911096765571e-07, "loss": 0.0131, "num_tokens": 1267019276.0, "reward": 0.02734375, "reward_std": 0.056218504905700684, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 1303 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.68359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1976.0, "completions/mean_length": 1645.13671875, "completions/mean_terminated_length": 774.7531127929688, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.4451651446616028, "frac_reward_zero_std": 0.8125, "grad_norm": 16.14362860214143, "learning_rate": 7.116909194515831e-07, "loss": -0.007, "num_tokens": 1267937458.0, "reward": 0.0625, "reward_std": 0.077679343521595, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.2422981858253479, "step": 1304 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2015.0, "completions/mean_length": 1700.390625, "completions/mean_terminated_length": 812.0555419921875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.4455065289749936, "frac_reward_zero_std": 0.65625, "grad_norm": 7.464543098024644, "learning_rate": 7.111904997359229e-07, "loss": 0.0329, "num_tokens": 1268881210.0, "reward": 0.083984375, "reward_std": 0.13156922161579132, "rewards/accuracy_reward/mean": 0.083984375, "rewards/accuracy_reward/std": 0.2776356339454651, "step": 1305 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.75390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1902.0, "completions/mean_length": 1708.720703125, "completions/mean_terminated_length": 669.34130859375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.4458479132883844, "frac_reward_zero_std": 0.75, "grad_norm": 16.470060939648388, "learning_rate": 7.106898512398305e-07, "loss": 0.0142, "num_tokens": 1269831115.0, "reward": 0.03125, "reward_std": 0.0792168527841568, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 1306 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2026.0, "completions/mean_length": 1679.484375, "completions/mean_terminated_length": 660.6470336914062, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.4461892976017752, "frac_reward_zero_std": 0.78125, "grad_norm": 21.563141507416802, "learning_rate": 7.101889746738848e-07, "loss": 0.0104, "num_tokens": 1270771715.0, "reward": 0.03515625, "reward_std": 0.07477159798145294, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 1307 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.732421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2041.0, "completions/mean_length": 1713.544921875, "completions/mean_terminated_length": 798.065673828125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.446530681915166, "frac_reward_zero_std": 0.71875, "grad_norm": 15.133919563267934, "learning_rate": 7.096878707489885e-07, "loss": 0.0068, "num_tokens": 1271724682.0, "reward": 0.044921875, "reward_std": 0.09841620922088623, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 1308 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1942.0, "completions/mean_length": 1631.69921875, "completions/mean_terminated_length": 748.3292236328125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.4468720662285568, "frac_reward_zero_std": 0.8125, "grad_norm": 20.20640398866373, "learning_rate": 7.091865401763671e-07, "loss": 0.0157, "num_tokens": 1272634848.0, "reward": 0.0546875, "reward_std": 0.06783140450716019, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 1309 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.68359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2018.0, "completions/mean_length": 1628.115234375, "completions/mean_terminated_length": 721.2777709960938, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.4472134505419476, "frac_reward_zero_std": 0.875, "grad_norm": 18.52241559846939, "learning_rate": 7.086849836675675e-07, "loss": 0.0002, "num_tokens": 1273557403.0, "reward": 0.041015625, "reward_std": 0.05303792655467987, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 1310 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1965.0, "completions/mean_length": 1593.05078125, "completions/mean_terminated_length": 693.7325439453125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.4475548348553384, "frac_reward_zero_std": 0.71875, "grad_norm": 4.82997478769421, "learning_rate": 7.081832019344573e-07, "loss": 0.0249, "num_tokens": 1274459061.0, "reward": 0.052734375, "reward_std": 0.10020478814840317, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 1311 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.70703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1844.0, "completions/mean_length": 1653.830078125, "completions/mean_terminated_length": 702.5667114257812, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.4478962191687292, "frac_reward_zero_std": 0.6875, "grad_norm": 9.618872630089466, "learning_rate": 7.076811956892241e-07, "loss": 0.0337, "num_tokens": 1275388926.0, "reward": 0.05859375, "reward_std": 0.09765692055225372, "rewards/accuracy_reward/mean": 0.05859375, "rewards/accuracy_reward/std": 0.23509246110916138, "step": 1312 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.646484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1999.0, "completions/mean_length": 1535.37890625, "completions/mean_terminated_length": 597.9337158203125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.44823760348212, "frac_reward_zero_std": 0.84375, "grad_norm": 11.404140867041663, "learning_rate": 7.07178965644374e-07, "loss": 0.0353, "num_tokens": 1276247008.0, "reward": 0.03125, "reward_std": 0.05789501965045929, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 1313 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.716796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2044.0, "completions/mean_length": 1656.814453125, "completions/mean_terminated_length": 666.7103271484375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.4485789877955108, "frac_reward_zero_std": 0.78125, "grad_norm": 0.9436619798384245, "learning_rate": 7.066765125127305e-07, "loss": 0.0321, "num_tokens": 1277175169.0, "reward": 0.041015625, "reward_std": 0.07669496536254883, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 1314 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.732421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2043.0, "completions/mean_length": 1686.98828125, "completions/mean_terminated_length": 698.8175048828125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.4489203721089016, "frac_reward_zero_std": 0.84375, "grad_norm": 22.719546786519388, "learning_rate": 7.061738370074342e-07, "loss": 0.0078, "num_tokens": 1278128395.0, "reward": 0.013671875, "reward_std": 0.043847277760505676, "rewards/accuracy_reward/mean": 0.013671875, "rewards/accuracy_reward/std": 0.1162383034825325, "step": 1315 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.646484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1971.0, "completions/mean_length": 1573.080078125, "completions/mean_terminated_length": 704.5801391601562, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.4492617564222924, "frac_reward_zero_std": 0.75, "grad_norm": 11.504743793403346, "learning_rate": 7.056709398419407e-07, "loss": 0.0207, "num_tokens": 1279016676.0, "reward": 0.06640625, "reward_std": 0.09338457882404327, "rewards/accuracy_reward/mean": 0.06640625, "rewards/accuracy_reward/std": 0.2492343932390213, "step": 1316 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.666015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2018.0, "completions/mean_length": 1611.3515625, "completions/mean_terminated_length": 740.6082153320312, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.4496031407356832, "frac_reward_zero_std": 0.8125, "grad_norm": 10.058086495513347, "learning_rate": 7.051678217300207e-07, "loss": 0.0247, "num_tokens": 1279914088.0, "reward": 0.052734375, "reward_std": 0.07680703699588776, "rewards/accuracy_reward/mean": 0.05443548411130905, "rewards/accuracy_reward/std": 0.227104052901268, "step": 1317 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.72265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2045.0, "completions/mean_length": 1701.88671875, "completions/mean_terminated_length": 800.042236328125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.449944525049074, "frac_reward_zero_std": 0.75, "grad_norm": 10.49812497251133, "learning_rate": 7.046644833857583e-07, "loss": 0.0027, "num_tokens": 1280858366.0, "reward": 0.03515625, "reward_std": 0.08120594918727875, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 1318 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2007.0, "completions/mean_length": 1650.828125, "completions/mean_terminated_length": 744.4615478515625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.4502859093624648, "frac_reward_zero_std": 0.71875, "grad_norm": 16.85734239284047, "learning_rate": 7.041609255235503e-07, "loss": 0.0239, "num_tokens": 1281783174.0, "reward": 0.0546875, "reward_std": 0.09429662674665451, "rewards/accuracy_reward/mean": 0.05645161122083664, "rewards/accuracy_reward/std": 0.23102475702762604, "step": 1319 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.716796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2036.0, "completions/mean_length": 1664.6015625, "completions/mean_terminated_length": 694.2069091796875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.4506272936758556, "frac_reward_zero_std": 0.71875, "grad_norm": 26.857474296486128, "learning_rate": 7.036571488581049e-07, "loss": 0.031, "num_tokens": 1282703578.0, "reward": 0.072265625, "reward_std": 0.11286985874176025, "rewards/accuracy_reward/mean": 0.072265625, "rewards/accuracy_reward/std": 0.2591804563999176, "step": 1320 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.0, "completions/mean_length": 1607.98828125, "completions/mean_terminated_length": 525.7973022460938, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.4509686779892464, "frac_reward_zero_std": 0.65625, "grad_norm": 18.14181889256335, "learning_rate": 7.031531541044411e-07, "loss": 0.0265, "num_tokens": 1283608340.0, "reward": 0.0546875, "reward_std": 0.11857913434505463, "rewards/accuracy_reward/mean": 0.05645161122083664, "rewards/accuracy_reward/std": 0.23102475702762604, "step": 1321 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.720703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1967.0, "completions/mean_length": 1636.759765625, "completions/mean_terminated_length": 575.58740234375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.4513100623026372, "frac_reward_zero_std": 0.875, "grad_norm": 3.0751683436281185, "learning_rate": 7.026489419778871e-07, "loss": 0.0016, "num_tokens": 1284518873.0, "reward": 0.03515625, "reward_std": 0.05012226477265358, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 1322 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.76953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1760.0, "completions/mean_length": 1737.1484375, "completions/mean_terminated_length": 699.2203369140625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.451651446616028, "frac_reward_zero_std": 0.8125, "grad_norm": 53.06411160893595, "learning_rate": 7.021445131940797e-07, "loss": 0.0, "num_tokens": 1285494389.0, "reward": 0.029296875, "reward_std": 0.06403100490570068, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 1323 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.70703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1814.0, "completions/mean_length": 1619.349609375, "completions/mean_terminated_length": 584.8733520507812, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.4519928309294188, "frac_reward_zero_std": 0.6875, "grad_norm": 28.868883303465935, "learning_rate": 7.016398684689636e-07, "loss": 0.017, "num_tokens": 1286394472.0, "reward": 0.087890625, "reward_std": 0.12906768918037415, "rewards/accuracy_reward/mean": 0.087890625, "rewards/accuracy_reward/std": 0.2834126651287079, "step": 1324 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.736328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1998.0, "completions/mean_length": 1682.6015625, "completions/mean_terminated_length": 662.1925659179688, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.4523342152428096, "frac_reward_zero_std": 0.875, "grad_norm": 5.302861087033077, "learning_rate": 7.011350085187895e-07, "loss": -0.0014, "num_tokens": 1287336220.0, "reward": 0.017578125, "reward_std": 0.04219770431518555, "rewards/accuracy_reward/mean": 0.018145160749554634, "rewards/accuracy_reward/std": 0.1336110234260559, "step": 1325 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.619140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1850.0, "completions/mean_length": 1486.162109375, "completions/mean_terminated_length": 572.8154296875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.4526755995562004, "frac_reward_zero_std": 0.6875, "grad_norm": 37.381653646933536, "learning_rate": 7.006299340601136e-07, "loss": 0.0037, "num_tokens": 1288173855.0, "reward": 0.0859375, "reward_std": 0.12120094895362854, "rewards/accuracy_reward/mean": 0.08870967477560043, "rewards/accuracy_reward/std": 0.2846112847328186, "step": 1326 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.712890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1781.0, "completions/mean_length": 1634.74609375, "completions/mean_terminated_length": 608.6394653320312, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.4530169838695912, "frac_reward_zero_std": 0.71875, "grad_norm": 11.249685721107372, "learning_rate": 7.001246458097972e-07, "loss": -0.0004, "num_tokens": 1289088509.0, "reward": 0.044921875, "reward_std": 0.0879673957824707, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 1327 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1928.0, "completions/mean_length": 1592.251953125, "completions/mean_terminated_length": 659.3511962890625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.453358368182982, "frac_reward_zero_std": 0.75, "grad_norm": 21.61853202556118, "learning_rate": 6.99619144485004e-07, "loss": 0.0037, "num_tokens": 1289987838.0, "reward": 0.048828125, "reward_std": 0.09363143146038055, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 1328 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2031.0, "completions/mean_length": 1649.35546875, "completions/mean_terminated_length": 668.9053955078125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.4536997524963728, "frac_reward_zero_std": 0.8125, "grad_norm": 12.049022847974536, "learning_rate": 6.99113430803201e-07, "loss": 0.0048, "num_tokens": 1290909444.0, "reward": 0.04296875, "reward_std": 0.06783141195774078, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 1329 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.701171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1881.0, "completions/mean_length": 1626.986328125, "completions/mean_terminated_length": 639.11767578125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.4540411368097636, "frac_reward_zero_std": 0.75, "grad_norm": 17.939744580731258, "learning_rate": 6.986075054821561e-07, "loss": 0.0274, "num_tokens": 1291806573.0, "reward": 0.046875, "reward_std": 0.09228022396564484, "rewards/accuracy_reward/mean": 0.046875, "rewards/accuracy_reward/std": 0.21157780289649963, "step": 1330 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.716796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2031.0, "completions/mean_length": 1673.810546875, "completions/mean_terminated_length": 726.72412109375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.4543825211231544, "frac_reward_zero_std": 0.75, "grad_norm": 25.076488254669922, "learning_rate": 6.98101369239938e-07, "loss": 0.0161, "num_tokens": 1292745724.0, "reward": 0.076171875, "reward_std": 0.09263914823532104, "rewards/accuracy_reward/mean": 0.076171875, "rewards/accuracy_reward/std": 0.26553234457969666, "step": 1331 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.74609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1675.0, "completions/mean_length": 1660.619140625, "completions/mean_terminated_length": 522.3153686523438, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.4547239054365452, "frac_reward_zero_std": 0.9375, "grad_norm": 8.475907272977782, "learning_rate": 6.975950227949143e-07, "loss": 0.0121, "num_tokens": 1293678937.0, "reward": 0.013671875, "reward_std": 0.026572702452540398, "rewards/accuracy_reward/mean": 0.013671875, "rewards/accuracy_reward/std": 0.1162383034825325, "step": 1332 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.69140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2030.0, "completions/mean_length": 1602.296875, "completions/mean_terminated_length": 603.6962280273438, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.455065289749936, "frac_reward_zero_std": 0.84375, "grad_norm": 32.3955670322673, "learning_rate": 6.970884668657512e-07, "loss": -0.0106, "num_tokens": 1294582545.0, "reward": 0.033203125, "reward_std": 0.059472277760505676, "rewards/accuracy_reward/mean": 0.033203125, "rewards/accuracy_reward/std": 0.17934183776378632, "step": 1333 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.755859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1996.0, "completions/mean_length": 1698.1953125, "completions/mean_terminated_length": 615.2000122070312, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.4554066740633268, "frac_reward_zero_std": 0.875, "grad_norm": 9.480896150343732, "learning_rate": 6.965817021714124e-07, "loss": -0.0002, "num_tokens": 1295527477.0, "reward": 0.021484375, "reward_std": 0.04456022381782532, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 1334 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.75, "completions/max_length": 2048.0, "completions/max_terminated_length": 2043.0, "completions/mean_length": 1702.421875, "completions/mean_terminated_length": 665.6875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.4557480583767176, "frac_reward_zero_std": 0.75, "grad_norm": 11.429878493826138, "learning_rate": 6.960747294311575e-07, "loss": -0.0045, "num_tokens": 1296471421.0, "reward": 0.0625, "reward_std": 0.10045164078474045, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.2422981858253479, "step": 1335 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.73046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1883.0, "completions/mean_length": 1659.87890625, "completions/mean_terminated_length": 608.0145263671875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.45608944269010837, "frac_reward_zero_std": 0.78125, "grad_norm": 58.035079576206755, "learning_rate": 6.955675493645415e-07, "loss": 0.0144, "num_tokens": 1297403887.0, "reward": 0.041015625, "reward_std": 0.08427447080612183, "rewards/accuracy_reward/mean": 0.04233871027827263, "rewards/accuracy_reward/std": 0.2015640139579773, "step": 1336 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.728515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2006.0, "completions/mean_length": 1678.25, "completions/mean_terminated_length": 686.0431518554688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.4564308270034992, "frac_reward_zero_std": 0.875, "grad_norm": 7.882906078572776, "learning_rate": 6.950601626914139e-07, "loss": 0.0005, "num_tokens": 1298342527.0, "reward": 0.03125, "reward_std": 0.04995594918727875, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 1337 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.654296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1993.0, "completions/mean_length": 1556.140625, "completions/mean_terminated_length": 625.2203369140625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.45677221131689, "frac_reward_zero_std": 0.75, "grad_norm": 20.07230014819777, "learning_rate": 6.94552570131917e-07, "loss": -0.0104, "num_tokens": 1299215991.0, "reward": 0.07421875, "reward_std": 0.09319227933883667, "rewards/accuracy_reward/mean": 0.07421875, "rewards/accuracy_reward/std": 0.2623828947544098, "step": 1338 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1855.0, "completions/mean_length": 1705.6015625, "completions/mean_terminated_length": 634.2257690429688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.4571135956302808, "frac_reward_zero_std": 0.65625, "grad_norm": 16.120989508695928, "learning_rate": 6.940447724064861e-07, "loss": 0.0116, "num_tokens": 1300157403.0, "reward": 0.068359375, "reward_std": 0.12741811573505402, "rewards/accuracy_reward/mean": 0.068359375, "rewards/accuracy_reward/std": 0.25260838866233826, "step": 1339 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1871.0, "completions/mean_length": 1659.267578125, "completions/mean_terminated_length": 540.1893920898438, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.45745497994367157, "frac_reward_zero_std": 0.8125, "grad_norm": 8.133899713613234, "learning_rate": 6.935367702358469e-07, "loss": 0.0127, "num_tokens": 1301089060.0, "reward": 0.0390625, "reward_std": 0.0736672431230545, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 1340 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1893.0, "completions/mean_length": 1621.078125, "completions/mean_terminated_length": 646.8204956054688, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.4577963642570624, "frac_reward_zero_std": 0.90625, "grad_norm": 5.193322477757516, "learning_rate": 6.930285643410154e-07, "loss": 0.0033, "num_tokens": 1301999420.0, "reward": 0.037109375, "reward_std": 0.04251687601208687, "rewards/accuracy_reward/mean": 0.03958333283662796, "rewards/accuracy_reward/std": 0.19518160820007324, "step": 1341 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.72265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1841.0, "completions/mean_length": 1655.626953125, "completions/mean_terminated_length": 633.2464599609375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.4581377485704532, "frac_reward_zero_std": 0.65625, "grad_norm": 26.415554485636136, "learning_rate": 6.925201554432972e-07, "loss": 0.023, "num_tokens": 1302932205.0, "reward": 0.080078125, "reward_std": 0.11682207137346268, "rewards/accuracy_reward/mean": 0.080078125, "rewards/accuracy_reward/std": 0.271679550409317, "step": 1342 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.693359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1999.0, "completions/mean_length": 1638.11328125, "completions/mean_terminated_length": 711.2993774414062, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.458479132883844, "frac_reward_zero_std": 0.78125, "grad_norm": 21.56890586359392, "learning_rate": 6.920115442642858e-07, "loss": 0.0177, "num_tokens": 1303844007.0, "reward": 0.03125, "reward_std": 0.0761418342590332, "rewards/accuracy_reward/mean": 0.032258063554763794, "rewards/accuracy_reward/std": 0.17686307430267334, "step": 1343 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.65625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2043.0, "completions/mean_length": 1587.7421875, "completions/mean_terminated_length": 709.0681762695312, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.45882051719723477, "frac_reward_zero_std": 0.78125, "grad_norm": 14.346002765247002, "learning_rate": 6.915027315258614e-07, "loss": 0.0167, "num_tokens": 1304733843.0, "reward": 0.056640625, "reward_std": 0.09118282794952393, "rewards/accuracy_reward/mean": 0.058467742055654526, "rewards/accuracy_reward/std": 0.23486268520355225, "step": 1344 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.666015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2034.0, "completions/mean_length": 1583.513671875, "completions/mean_terminated_length": 657.25732421875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.4591619015106256, "frac_reward_zero_std": 0.84375, "grad_norm": 5.59197392701996, "learning_rate": 6.909937179501908e-07, "loss": 0.0097, "num_tokens": 1305615290.0, "reward": 0.072265625, "reward_std": 0.06733252108097076, "rewards/accuracy_reward/mean": 0.07459677755832672, "rewards/accuracy_reward/std": 0.263004869222641, "step": 1345 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1924.0, "completions/mean_length": 1540.7578125, "completions/mean_terminated_length": 636.5435180664062, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.4595032858240164, "frac_reward_zero_std": 0.875, "grad_norm": 10.76918297692418, "learning_rate": 6.904845042597258e-07, "loss": 0.008, "num_tokens": 1306476302.0, "reward": 0.02734375, "reward_std": 0.04847269132733345, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 1346 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2034.0, "completions/mean_length": 1648.04296875, "completions/mean_terminated_length": 645.8013916015625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.4598446701374072, "frac_reward_zero_std": 0.71875, "grad_norm": 18.55915123733681, "learning_rate": 6.899750911772019e-07, "loss": 0.0299, "num_tokens": 1307407716.0, "reward": 0.0625, "reward_std": 0.10355263948440552, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.2422981858253479, "step": 1347 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.630859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1996.0, "completions/mean_length": 1508.560546875, "completions/mean_terminated_length": 586.661376953125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.46018605445079797, "frac_reward_zero_std": 0.8125, "grad_norm": 26.12937549356265, "learning_rate": 6.894654794256378e-07, "loss": -0.0002, "num_tokens": 1308253811.0, "reward": 0.037109375, "reward_std": 0.06166848540306091, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 1348 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.732421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2029.0, "completions/mean_length": 1654.640625, "completions/mean_terminated_length": 577.927001953125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.4605274387641888, "frac_reward_zero_std": 0.71875, "grad_norm": 20.454096070937826, "learning_rate": 6.889556697283344e-07, "loss": 0.0083, "num_tokens": 1309182027.0, "reward": 0.041015625, "reward_std": 0.09792976081371307, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 1349 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.755859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1779.0, "completions/mean_length": 1706.724609375, "completions/mean_terminated_length": 650.6160278320312, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.4608688230775796, "frac_reward_zero_std": 0.84375, "grad_norm": 7.550627450987065, "learning_rate": 6.88445662808873e-07, "loss": 0.0072, "num_tokens": 1310135918.0, "reward": 0.013671875, "reward_std": 0.04478531330823898, "rewards/accuracy_reward/mean": 0.013671875, "rewards/accuracy_reward/std": 0.1162383034825325, "step": 1350 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2008.0, "completions/mean_length": 1543.162109375, "completions/mean_terminated_length": 643.2337036132812, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.4612102073909704, "frac_reward_zero_std": 0.8125, "grad_norm": 6.1865046834519575, "learning_rate": 6.879354593911154e-07, "loss": 0.005, "num_tokens": 1311001233.0, "reward": 0.037109375, "reward_std": 0.06602106243371964, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 1351 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1703.0, "completions/mean_length": 1669.40625, "completions/mean_terminated_length": 622.7058715820312, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.46155159170436116, "frac_reward_zero_std": 0.71875, "grad_norm": 10.561579161379623, "learning_rate": 6.874250601992019e-07, "loss": 0.0119, "num_tokens": 1311933457.0, "reward": 0.056640625, "reward_std": 0.09778349101543427, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 1352 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.583984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1936.0, "completions/mean_length": 1464.619140625, "completions/mean_terminated_length": 645.69482421875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.461892976017752, "frac_reward_zero_std": 0.875, "grad_norm": 60.39423560879323, "learning_rate": 6.869144659575507e-07, "loss": -0.0078, "num_tokens": 1312769870.0, "reward": 0.03515625, "reward_std": 0.052318472415208817, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 1353 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1986.0, "completions/mean_length": 1634.63671875, "completions/mean_terminated_length": 655.618408203125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.4622343603311428, "frac_reward_zero_std": 0.84375, "grad_norm": 43.96253666820117, "learning_rate": 6.864036773908572e-07, "loss": 0.0219, "num_tokens": 1313689620.0, "reward": 0.05078125, "reward_std": 0.06956121325492859, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 1354 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.654296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1962.0, "completions/mean_length": 1558.298828125, "completions/mean_terminated_length": 631.4632568359375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.4625757446445336, "frac_reward_zero_std": 0.84375, "grad_norm": 11.085300261523463, "learning_rate": 6.858926952240925e-07, "loss": 0.0185, "num_tokens": 1314561437.0, "reward": 0.0625, "reward_std": 0.06464291363954544, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.2422981858253479, "step": 1355 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.619140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1941.0, "completions/mean_length": 1494.4921875, "completions/mean_terminated_length": 594.6871948242188, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.46291712895792436, "frac_reward_zero_std": 0.875, "grad_norm": 8.37901856552808, "learning_rate": 6.853815201825016e-07, "loss": 0.0105, "num_tokens": 1315407065.0, "reward": 0.015625, "reward_std": 0.04081955552101135, "rewards/accuracy_reward/mean": 0.015625, "rewards/accuracy_reward/std": 0.12414088100194931, "step": 1356 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.564453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1970.0, "completions/mean_length": 1451.478515625, "completions/mean_terminated_length": 678.4080810546875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.4632585132713152, "frac_reward_zero_std": 0.78125, "grad_norm": 17.749692656207962, "learning_rate": 6.848701529916047e-07, "loss": -0.0162, "num_tokens": 1316222974.0, "reward": 0.0390625, "reward_std": 0.07807311415672302, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 1357 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.626953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1974.0, "completions/mean_length": 1516.451171875, "completions/mean_terminated_length": 623.1151733398438, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.463599897584706, "frac_reward_zero_std": 0.78125, "grad_norm": 11.6380802932008, "learning_rate": 6.843585943771935e-07, "loss": 0.013, "num_tokens": 1317083957.0, "reward": 0.064453125, "reward_std": 0.08973139524459839, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 1358 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1978.0, "completions/mean_length": 1579.03125, "completions/mean_terminated_length": 619.0654907226562, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.4639412818980968, "frac_reward_zero_std": 0.65625, "grad_norm": 15.0739240447066, "learning_rate": 6.838468450653322e-07, "loss": 0.0082, "num_tokens": 1317977877.0, "reward": 0.041015625, "reward_std": 0.10805703699588776, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 1359 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.64453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2048.0, "completions/mean_length": 1548.1171875, "completions/mean_terminated_length": 641.7362670898438, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.46428266621148756, "frac_reward_zero_std": 0.75, "grad_norm": 13.239727925845521, "learning_rate": 6.833349057823553e-07, "loss": -0.0137, "num_tokens": 1318871617.0, "reward": 0.02734375, "reward_std": 0.07779236882925034, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 1360 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.568359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2032.0, "completions/mean_length": 1439.478515625, "completions/mean_terminated_length": 638.4072875976562, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.4646240505248784, "frac_reward_zero_std": 0.65625, "grad_norm": 20.280329262494572, "learning_rate": 6.828227772548669e-07, "loss": 0.0145, "num_tokens": 1319689558.0, "reward": 0.07421875, "reward_std": 0.12922608852386475, "rewards/accuracy_reward/mean": 0.07421875, "rewards/accuracy_reward/std": 0.2623828947544098, "step": 1361 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.66015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2017.0, "completions/mean_length": 1573.74609375, "completions/mean_terminated_length": 652.4942626953125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.4649654348382692, "frac_reward_zero_std": 0.75, "grad_norm": 9.62762146773623, "learning_rate": 6.823104602097398e-07, "loss": 0.0225, "num_tokens": 1320579236.0, "reward": 0.041015625, "reward_std": 0.08611729741096497, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 1362 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.646484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1961.0, "completions/mean_length": 1576.84375, "completions/mean_terminated_length": 715.4364624023438, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.46530681915166, "frac_reward_zero_std": 0.90625, "grad_norm": 3.6505834996527624, "learning_rate": 6.817979553741143e-07, "loss": 0.0104, "num_tokens": 1321467924.0, "reward": 0.0234375, "reward_std": 0.036547206342220306, "rewards/accuracy_reward/mean": 0.024193547666072845, "rewards/accuracy_reward/std": 0.15380479395389557, "step": 1363 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.62890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2039.0, "completions/mean_length": 1538.275390625, "completions/mean_terminated_length": 674.4263305664062, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.46564820346505076, "frac_reward_zero_std": 0.84375, "grad_norm": 13.193890897779049, "learning_rate": 6.812852634753974e-07, "loss": 0.01, "num_tokens": 1322335809.0, "reward": 0.03125, "reward_std": 0.05914659798145294, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 1364 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.751953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1933.0, "completions/mean_length": 1724.220703125, "completions/mean_terminated_length": 742.68505859375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.46598958777844157, "frac_reward_zero_std": 0.8125, "grad_norm": 9.945393750644513, "learning_rate": 6.807723852412613e-07, "loss": 0.0037, "num_tokens": 1323301506.0, "reward": 0.04296875, "reward_std": 0.07289456576108932, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 1365 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.66015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2032.0, "completions/mean_length": 1579.826171875, "completions/mean_terminated_length": 670.3850708007812, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.4663309720918324, "frac_reward_zero_std": 0.75, "grad_norm": 7.553615652351151, "learning_rate": 6.802593213996431e-07, "loss": 0.0101, "num_tokens": 1324185017.0, "reward": 0.064453125, "reward_std": 0.09488959610462189, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 1366 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.736328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2019.0, "completions/mean_length": 1680.423828125, "completions/mean_terminated_length": 653.933349609375, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.4666723564052232, "frac_reward_zero_std": 0.78125, "grad_norm": 9.653907535189768, "learning_rate": 6.797460726787427e-07, "loss": 0.0119, "num_tokens": 1325115810.0, "reward": 0.0546875, "reward_std": 0.08464758098125458, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 1367 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.68359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1979.0, "completions/mean_length": 1637.015625, "completions/mean_terminated_length": 749.08642578125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.46701374071861396, "frac_reward_zero_std": 0.90625, "grad_norm": 2.9808912382522084, "learning_rate": 6.792326398070233e-07, "loss": 0.0087, "num_tokens": 1326026442.0, "reward": 0.021484375, "reward_std": 0.036420635879039764, "rewards/accuracy_reward/mean": 0.02217741869390011, "rewards/accuracy_reward/std": 0.14740893244743347, "step": 1368 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.66796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1988.0, "completions/mean_length": 1617.9140625, "completions/mean_terminated_length": 752.682373046875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.46735512503200477, "frac_reward_zero_std": 0.6875, "grad_norm": 17.93567664268745, "learning_rate": 6.787190235132085e-07, "loss": 0.0152, "num_tokens": 1326932206.0, "reward": 0.029296875, "reward_std": 0.09055596590042114, "rewards/accuracy_reward/mean": 0.030241934582591057, "rewards/accuracy_reward/std": 0.1714252382516861, "step": 1369 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.689453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1985.0, "completions/mean_length": 1601.900390625, "completions/mean_terminated_length": 611.5031127929688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.4676965093453956, "frac_reward_zero_std": 0.6875, "grad_norm": 12.076609858561431, "learning_rate": 6.782052245262829e-07, "loss": 0.0111, "num_tokens": 1327837003.0, "reward": 0.080078125, "reward_std": 0.12356996536254883, "rewards/accuracy_reward/mean": 0.080078125, "rewards/accuracy_reward/std": 0.271679550409317, "step": 1370 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1783.0, "completions/mean_length": 1588.802734375, "completions/mean_terminated_length": 578.5687866210938, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.4680378936587864, "frac_reward_zero_std": 0.75, "grad_norm": 4.336793567182062, "learning_rate": 6.7769124357549e-07, "loss": -0.0082, "num_tokens": 1328720662.0, "reward": 0.05078125, "reward_std": 0.09028453379869461, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 1371 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1898.0, "completions/mean_length": 1754.19140625, "completions/mean_terminated_length": 834.8547973632812, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.46837927797217715, "frac_reward_zero_std": 0.75, "grad_norm": 19.450924548088405, "learning_rate": 6.77177081390332e-07, "loss": 0.0155, "num_tokens": 1329686808.0, "reward": 0.056640625, "reward_std": 0.08039448410272598, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 1372 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.740234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1843.0, "completions/mean_length": 1676.798828125, "completions/mean_terminated_length": 619.0150756835938, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.46872066228556797, "frac_reward_zero_std": 0.875, "grad_norm": 10.423625513619431, "learning_rate": 6.76662738700568e-07, "loss": -0.0157, "num_tokens": 1330616897.0, "reward": 0.03515625, "reward_std": 0.04554459825158119, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 1373 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.759765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2040.0, "completions/mean_length": 1698.12890625, "completions/mean_terminated_length": 591.6259765625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.4690620465989588, "frac_reward_zero_std": 0.875, "grad_norm": 0.9431566818651097, "learning_rate": 6.761482162362134e-07, "loss": 0.015, "num_tokens": 1331561891.0, "reward": 0.03515625, "reward_std": 0.04505910724401474, "rewards/accuracy_reward/mean": 0.03629032149910927, "rewards/accuracy_reward/std": 0.1872003972530365, "step": 1374 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2043.0, "completions/mean_length": 1668.2734375, "completions/mean_terminated_length": 659.2857055664062, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.4694034309123496, "frac_reward_zero_std": 0.8125, "grad_norm": 13.829856259882128, "learning_rate": 6.756335147275387e-07, "loss": 0.0074, "num_tokens": 1332497279.0, "reward": 0.052734375, "reward_std": 0.06092274561524391, "rewards/accuracy_reward/mean": 0.05443548411130905, "rewards/accuracy_reward/std": 0.2271040380001068, "step": 1375 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.77734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1973.0, "completions/mean_length": 1717.93359375, "completions/mean_terminated_length": 565.5964965820312, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.46974481522574035, "frac_reward_zero_std": 0.875, "grad_norm": 11.582070368057904, "learning_rate": 6.751186349050683e-07, "loss": -0.0008, "num_tokens": 1333454317.0, "reward": 0.03125, "reward_std": 0.05012226477265358, "rewards/accuracy_reward/mean": 0.032258063554763794, "rewards/accuracy_reward/std": 0.17686307430267334, "step": 1376 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1863.0, "completions/mean_length": 1746.841796875, "completions/mean_terminated_length": 506.0699768066406, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.47008619953913117, "frac_reward_zero_std": 0.84375, "grad_norm": 20.412395649159794, "learning_rate": 6.746035774995805e-07, "loss": 0.0144, "num_tokens": 1334422444.0, "reward": 0.056640625, "reward_std": 0.061668481677770615, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 1377 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.837890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1923.0, "completions/mean_length": 1836.814453125, "completions/mean_terminated_length": 745.2650146484375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.470427583852522, "frac_reward_zero_std": 0.96875, "grad_norm": 1.6495096115336052, "learning_rate": 6.740883432421044e-07, "loss": -0.0012, "num_tokens": 1335443693.0, "reward": 0.001953125, "reward_std": 0.0078125, "rewards/accuracy_reward/mean": 0.002016128972172737, "rewards/accuracy_reward/std": 0.044901326298713684, "step": 1378 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2008.0, "completions/mean_length": 1672.365234375, "completions/mean_terminated_length": 730.7054443359375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.4707689681659128, "frac_reward_zero_std": 0.8125, "grad_norm": 8.260373586386889, "learning_rate": 6.735729328639213e-07, "loss": -0.0028, "num_tokens": 1336374472.0, "reward": 0.048828125, "reward_std": 0.05782270431518555, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 1379 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.84765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1898.0, "completions/mean_length": 1839.603515625, "completions/mean_terminated_length": 680.0640869140625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.47111035247930355, "frac_reward_zero_std": 0.90625, "grad_norm": 2.829318815390029, "learning_rate": 6.730573470965618e-07, "loss": -0.0047, "num_tokens": 1337406621.0, "reward": 0.009765625, "reward_std": 0.029160313308238983, "rewards/accuracy_reward/mean": 0.009765625, "rewards/accuracy_reward/std": 0.09843364357948303, "step": 1380 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1820.0, "completions/mean_length": 1780.373046875, "completions/mean_terminated_length": 779.6018676757812, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.47145173679269436, "frac_reward_zero_std": 0.84375, "grad_norm": 17.86707798643111, "learning_rate": 6.725415866718055e-07, "loss": 0.011, "num_tokens": 1338402748.0, "reward": 0.0546875, "reward_std": 0.06441686302423477, "rewards/accuracy_reward/mean": 0.05645161122083664, "rewards/accuracy_reward/std": 0.23102475702762604, "step": 1381 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.775390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1891.0, "completions/mean_length": 1742.970703125, "completions/mean_terminated_length": 689.9564819335938, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.4717931211060852, "frac_reward_zero_std": 0.9375, "grad_norm": 6.206031883431773, "learning_rate": 6.720256523216802e-07, "loss": -0.0027, "num_tokens": 1339369949.0, "reward": 0.00390625, "reward_std": 0.015625, "rewards/accuracy_reward/mean": 0.00390625, "rewards/accuracy_reward/std": 0.06243881583213806, "step": 1382 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.82421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1887.0, "completions/mean_length": 1792.486328125, "completions/mean_terminated_length": 594.4111328125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.472134505419476, "frac_reward_zero_std": 0.8125, "grad_norm": 2.040164009486912, "learning_rate": 6.715095447784602e-07, "loss": 0.0111, "num_tokens": 1340358086.0, "reward": 0.017578125, "reward_std": 0.05452118441462517, "rewards/accuracy_reward/mean": 0.017578125, "rewards/accuracy_reward/std": 0.13154059648513794, "step": 1383 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2042.0, "completions/mean_length": 1858.412109375, "completions/mean_terminated_length": 699.8194580078125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.47247588973286675, "frac_reward_zero_std": 0.8125, "grad_norm": 6.846622189400305, "learning_rate": 6.709932647746659e-07, "loss": 0.0299, "num_tokens": 1341384425.0, "reward": 0.064453125, "reward_std": 0.07296784222126007, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 1384 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.837890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2031.0, "completions/mean_length": 1820.94921875, "completions/mean_terminated_length": 647.3975830078125, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.47281727404625756, "frac_reward_zero_std": 0.84375, "grad_norm": 3.3114748835495895, "learning_rate": 6.704768130430619e-07, "loss": 0.016, "num_tokens": 1342401855.0, "reward": 0.013671875, "reward_std": 0.04478531330823898, "rewards/accuracy_reward/mean": 0.013671875, "rewards/accuracy_reward/std": 0.1162383034825325, "step": 1385 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.791015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1886.0, "completions/mean_length": 1749.419921875, "completions/mean_terminated_length": 619.2803344726562, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.4731586583596484, "frac_reward_zero_std": 0.84375, "grad_norm": 7.1082993943323824, "learning_rate": 6.699601903166575e-07, "loss": 0.0215, "num_tokens": 1343372294.0, "reward": 0.041015625, "reward_std": 0.06508206576108932, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 1386 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.798828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1929.0, "completions/mean_length": 1768.927734375, "completions/mean_terminated_length": 660.7669677734375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.4735000426730392, "frac_reward_zero_std": 0.84375, "grad_norm": 3.2609648743078488, "learning_rate": 6.69443397328704e-07, "loss": 0.0163, "num_tokens": 1344356241.0, "reward": 0.029296875, "reward_std": 0.060957904905080795, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 1387 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1800.0, "completions/mean_length": 1699.99609375, "completions/mean_terminated_length": 563.183349609375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.47384142698642995, "frac_reward_zero_std": 0.78125, "grad_norm": 14.579897856451089, "learning_rate": 6.689264348126944e-07, "loss": 0.0368, "num_tokens": 1345300383.0, "reward": 0.0625, "reward_std": 0.0892922431230545, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.2422981858253479, "step": 1388 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.75390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1945.0, "completions/mean_length": 1686.861328125, "completions/mean_terminated_length": 580.5159301757812, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.47418281129982076, "frac_reward_zero_std": 0.84375, "grad_norm": 5.267659316297761, "learning_rate": 6.684093035023626e-07, "loss": 0.0113, "num_tokens": 1346246504.0, "reward": 0.05859375, "reward_std": 0.06640692055225372, "rewards/accuracy_reward/mean": 0.060483869165182114, "rewards/accuracy_reward/std": 0.2386218160390854, "step": 1389 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.79296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1910.0, "completions/mean_length": 1768.349609375, "completions/mean_terminated_length": 697.23583984375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.4745241956132116, "frac_reward_zero_std": 0.8125, "grad_norm": 25.18048385464495, "learning_rate": 6.678920041316818e-07, "loss": 0.0114, "num_tokens": 1347223643.0, "reward": 0.0625, "reward_std": 0.07284127175807953, "rewards/accuracy_reward/mean": 0.06451612710952759, "rewards/accuracy_reward/std": 0.24591811001300812, "step": 1390 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.826171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1963.0, "completions/mean_length": 1809.603515625, "completions/mean_terminated_length": 676.5505981445312, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.4748655799266024, "frac_reward_zero_std": 0.875, "grad_norm": 21.528063056077638, "learning_rate": 6.67374537434864e-07, "loss": -0.0036, "num_tokens": 1348233552.0, "reward": 0.013671875, "reward_std": 0.03889618441462517, "rewards/accuracy_reward/mean": 0.013671875, "rewards/accuracy_reward/std": 0.1162383034825325, "step": 1391 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.76953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2001.0, "completions/mean_length": 1756.994140625, "completions/mean_terminated_length": 785.3305053710938, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.47520696423999315, "frac_reward_zero_std": 0.90625, "grad_norm": 14.381690105779146, "learning_rate": 6.668569041463582e-07, "loss": 0.0021, "num_tokens": 1349212429.0, "reward": 0.017578125, "reward_std": 0.03630761057138443, "rewards/accuracy_reward/mean": 0.017578125, "rewards/accuracy_reward/std": 0.13154059648513794, "step": 1392 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.79296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1925.0, "completions/mean_length": 1771.447265625, "completions/mean_terminated_length": 712.1981201171875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.47554834855338396, "frac_reward_zero_std": 0.75, "grad_norm": 49.66768913374407, "learning_rate": 6.663391050008505e-07, "loss": 0.0036, "num_tokens": 1350201090.0, "reward": 0.06640625, "reward_std": 0.09891509264707565, "rewards/accuracy_reward/mean": 0.06640625, "rewards/accuracy_reward/std": 0.2492343932390213, "step": 1393 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.77734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1942.0, "completions/mean_length": 1766.05859375, "completions/mean_terminated_length": 781.7368774414062, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.4758897328667748, "frac_reward_zero_std": 0.75, "grad_norm": 22.757631971569335, "learning_rate": 6.658211407332619e-07, "loss": 0.0051, "num_tokens": 1351178896.0, "reward": 0.09765625, "reward_std": 0.10668125003576279, "rewards/accuracy_reward/mean": 0.09765625, "rewards/accuracy_reward/std": 0.29713961482048035, "step": 1394 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.81640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2034.0, "completions/mean_length": 1798.927734375, "completions/mean_terminated_length": 691.3510131835938, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.4762311171801656, "frac_reward_zero_std": 0.875, "grad_norm": 6.678833770275776, "learning_rate": 6.65303012078748e-07, "loss": 0.0157, "num_tokens": 1352179195.0, "reward": 0.015625, "reward_std": 0.03839729726314545, "rewards/accuracy_reward/mean": 0.015625, "rewards/accuracy_reward/std": 0.12414088100194931, "step": 1395 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.83984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1940.0, "completions/mean_length": 1825.376953125, "completions/mean_terminated_length": 658.695068359375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.47657250149355634, "frac_reward_zero_std": 0.8125, "grad_norm": 6.523591245166773, "learning_rate": 6.647847197726978e-07, "loss": 0.0158, "num_tokens": 1353190380.0, "reward": 0.01953125, "reward_std": 0.055899329483509064, "rewards/accuracy_reward/mean": 0.01953125, "rewards/accuracy_reward/std": 0.1385180652141571, "step": 1396 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.751953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1910.0, "completions/mean_length": 1704.1171875, "completions/mean_terminated_length": 661.6378173828125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.47691388580694716, "frac_reward_zero_std": 0.90625, "grad_norm": 1.41450783451255, "learning_rate": 6.642662645507322e-07, "loss": 0.0066, "num_tokens": 1354137128.0, "reward": 0.005859375, "reward_std": 0.0234375, "rewards/accuracy_reward/mean": 0.006048386916518211, "rewards/accuracy_reward/std": 0.07761410623788834, "step": 1397 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.802734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1908.0, "completions/mean_length": 1774.408203125, "completions/mean_terminated_length": 661.0792236328125, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.47725527012033797, "frac_reward_zero_std": 0.75, "grad_norm": 12.09313715087493, "learning_rate": 6.637476471487036e-07, "loss": 0.0152, "num_tokens": 1355117145.0, "reward": 0.021484375, "reward_std": 0.07108421623706818, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 1398 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.77734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1966.0, "completions/mean_length": 1766.380859375, "completions/mean_terminated_length": 783.5526123046875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.4775966544337288, "frac_reward_zero_std": 0.875, "grad_norm": 22.956806215144805, "learning_rate": 6.632288683026946e-07, "loss": 0.0087, "num_tokens": 1356108524.0, "reward": 0.02734375, "reward_std": 0.04517117515206337, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 1399 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2010.0, "completions/mean_length": 1682.267578125, "completions/mean_terminated_length": 747.6180419921875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.47793803874711954, "frac_reward_zero_std": 0.71875, "grad_norm": 11.14464243875765, "learning_rate": 6.62709928749017e-07, "loss": 0.0275, "num_tokens": 1357046837.0, "reward": 0.048828125, "reward_std": 0.09852826595306396, "rewards/accuracy_reward/mean": 0.05040322616696358, "rewards/accuracy_reward/std": 0.21899642050266266, "step": 1400 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.73828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2005.0, "completions/mean_length": 1719.34375, "completions/mean_terminated_length": 792.23876953125, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.47827942306051036, "frac_reward_zero_std": 0.84375, "grad_norm": 7.635424386954305, "learning_rate": 6.621908292242104e-07, "loss": 0.0018, "num_tokens": 1358002965.0, "reward": 0.044921875, "reward_std": 0.06292665004730225, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 1401 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2039.0, "completions/mean_length": 1732.890625, "completions/mean_terminated_length": 703.9166870117188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.47862080737390117, "frac_reward_zero_std": 0.9375, "grad_norm": 0.091214312445108, "learning_rate": 6.616715704650418e-07, "loss": 0.0052, "num_tokens": 1358973341.0, "reward": 0.01953125, "reward_std": 0.023823359981179237, "rewards/accuracy_reward/mean": 0.01953125, "rewards/accuracy_reward/std": 0.1385180652141571, "step": 1402 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.775390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2004.0, "completions/mean_length": 1744.53515625, "completions/mean_terminated_length": 696.9216918945312, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.478962191687292, "frac_reward_zero_std": 0.90625, "grad_norm": 66.85219103796454, "learning_rate": 6.611521532085038e-07, "loss": 0.0054, "num_tokens": 1359946463.0, "reward": 0.005859375, "reward_std": 0.0234375, "rewards/accuracy_reward/mean": 0.006048386916518211, "rewards/accuracy_reward/std": 0.07761410623788834, "step": 1403 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1742.2734375, "completions/mean_terminated_length": 785.6451416015625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.4793035760006828, "frac_reward_zero_std": 0.8125, "grad_norm": 12.18161866080606, "learning_rate": 6.606325781918144e-07, "loss": 0.0173, "num_tokens": 1360907803.0, "reward": 0.029296875, "reward_std": 0.06645326316356659, "rewards/accuracy_reward/mean": 0.030241934582591057, "rewards/accuracy_reward/std": 0.1714252382516861, "step": 1404 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.822265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2038.0, "completions/mean_length": 1841.951171875, "completions/mean_terminated_length": 888.6923217773438, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.47964496031407355, "frac_reward_zero_std": 0.875, "grad_norm": 2.0392390766528123, "learning_rate": 6.601128461524152e-07, "loss": 0.0215, "num_tokens": 1361930402.0, "reward": 0.01953125, "reward_std": 0.04505910724401474, "rewards/accuracy_reward/mean": 0.01953125, "rewards/accuracy_reward/std": 0.1385180652141571, "step": 1405 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.794921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1926.0, "completions/mean_length": 1764.896484375, "completions/mean_terminated_length": 667.5333862304688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.47998634462746437, "frac_reward_zero_std": 0.71875, "grad_norm": 26.706955843157182, "learning_rate": 6.595929578279708e-07, "loss": 0.0276, "num_tokens": 1362912893.0, "reward": 0.07421875, "reward_std": 0.11462032794952393, "rewards/accuracy_reward/mean": 0.07421875, "rewards/accuracy_reward/std": 0.2623828947544098, "step": 1406 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.78515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2022.0, "completions/mean_length": 1769.736328125, "completions/mean_terminated_length": 752.80908203125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.4803277289408552, "frac_reward_zero_std": 0.8125, "grad_norm": 23.363204228157326, "learning_rate": 6.590729139563675e-07, "loss": 0.0121, "num_tokens": 1363894006.0, "reward": 0.02734375, "reward_std": 0.06546888500452042, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 1407 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.783203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.0, "completions/mean_length": 1789.51953125, "completions/mean_terminated_length": 855.729736328125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.480669113254246, "frac_reward_zero_std": 0.84375, "grad_norm": 3.9388193938095077, "learning_rate": 6.585527152757128e-07, "loss": 0.0133, "num_tokens": 1364886080.0, "reward": 0.02734375, "reward_std": 0.05864076316356659, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 1408 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1786.0, "completions/mean_length": 1820.185546875, "completions/mean_terminated_length": 659.4166870117188, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.48101049756763675, "frac_reward_zero_std": 0.78125, "grad_norm": 9.61702704796192, "learning_rate": 6.580323625243332e-07, "loss": 0.0068, "num_tokens": 1365897167.0, "reward": 0.04296875, "reward_std": 0.07135801762342453, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 1409 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2028.0, "completions/mean_length": 1661.72265625, "completions/mean_terminated_length": 712.0811157226562, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.48135188188102757, "frac_reward_zero_std": 0.78125, "grad_norm": 53.40134156704168, "learning_rate": 6.575118564407742e-07, "loss": 0.0003, "num_tokens": 1366830945.0, "reward": 0.056640625, "reward_std": 0.0906703919172287, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 1410 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.83203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1984.0, "completions/mean_length": 1830.974609375, "completions/mean_terminated_length": 755.9418334960938, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.4816932661944184, "frac_reward_zero_std": 0.8125, "grad_norm": 28.88819429655891, "learning_rate": 6.569911977637994e-07, "loss": 0.0215, "num_tokens": 1367851524.0, "reward": 0.0234375, "reward_std": 0.05974511429667473, "rewards/accuracy_reward/mean": 0.0234375, "rewards/accuracy_reward/std": 0.15143637359142303, "step": 1411 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1978.0, "completions/mean_length": 1809.4375, "completions/mean_terminated_length": 720.3478393554688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.4820346505078092, "frac_reward_zero_std": 0.875, "grad_norm": 10.988399397620244, "learning_rate": 6.564703872323883e-07, "loss": 0.0124, "num_tokens": 1368857700.0, "reward": 0.01953125, "reward_std": 0.04318207502365112, "rewards/accuracy_reward/mean": 0.01953125, "rewards/accuracy_reward/std": 0.1385180652141571, "step": 1412 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.0, "completions/mean_length": 1733.482421875, "completions/mean_terminated_length": 659.7844848632812, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.48237603482119995, "frac_reward_zero_std": 0.8125, "grad_norm": 9.623516995992913, "learning_rate": 6.559494255857362e-07, "loss": 0.0106, "num_tokens": 1369818187.0, "reward": 0.037109375, "reward_std": 0.06613312661647797, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 1413 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.779296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1904.0, "completions/mean_length": 1742.388671875, "completions/mean_terminated_length": 663.283203125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.48271741913459076, "frac_reward_zero_std": 0.875, "grad_norm": 14.566109757620035, "learning_rate": 6.554283135632529e-07, "loss": 0.0047, "num_tokens": 1370787186.0, "reward": 0.01171875, "reward_std": 0.03697281330823898, "rewards/accuracy_reward/mean": 0.01171875, "rewards/accuracy_reward/std": 0.10772226005792618, "step": 1414 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1870.0, "completions/mean_length": 1730.39453125, "completions/mean_terminated_length": 646.1551513671875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.4830588034479816, "frac_reward_zero_std": 0.75, "grad_norm": 40.17480950569346, "learning_rate": 6.549070519045615e-07, "loss": 0.0198, "num_tokens": 1371747788.0, "reward": 0.056640625, "reward_std": 0.09761717915534973, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 1415 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1794.0, "completions/mean_length": 1803.20703125, "completions/mean_terminated_length": 685.6739501953125, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.4834001877613724, "frac_reward_zero_std": 0.71875, "grad_norm": 63.31516703249113, "learning_rate": 6.543856413494979e-07, "loss": 0.0217, "num_tokens": 1372750182.0, "reward": 0.0703125, "reward_std": 0.10540273785591125, "rewards/accuracy_reward/mean": 0.0703125, "rewards/accuracy_reward/std": 0.25592297315597534, "step": 1416 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.830078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2035.0, "completions/mean_length": 1832.626953125, "completions/mean_terminated_length": 781.0919799804688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.48374157207476315, "frac_reward_zero_std": 0.78125, "grad_norm": 21.988251705941856, "learning_rate": 6.538640826381086e-07, "loss": 0.0253, "num_tokens": 1373763191.0, "reward": 0.04296875, "reward_std": 0.08004186302423477, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 1417 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.841796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1922.0, "completions/mean_length": 1834.294921875, "completions/mean_terminated_length": 697.1728515625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.48408295638815396, "frac_reward_zero_std": 0.84375, "grad_norm": 18.109590450691698, "learning_rate": 6.533423765106509e-07, "loss": 0.0196, "num_tokens": 1374783614.0, "reward": 0.03515625, "reward_std": 0.06244811415672302, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 1418 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.763671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1759.0, "completions/mean_length": 1700.357421875, "completions/mean_terminated_length": 576.9834594726562, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.4844243407015448, "frac_reward_zero_std": 0.78125, "grad_norm": 19.07382527230103, "learning_rate": 6.528205237075916e-07, "loss": 0.0049, "num_tokens": 1375727893.0, "reward": 0.0390625, "reward_std": 0.07775957137346268, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 1419 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.810546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2014.0, "completions/mean_length": 1793.203125, "completions/mean_terminated_length": 703.0927734375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.4847657250149356, "frac_reward_zero_std": 0.78125, "grad_norm": 15.816900049044747, "learning_rate": 6.522985249696049e-07, "loss": 0.0253, "num_tokens": 1376722285.0, "reward": 0.03515625, "reward_std": 0.0744580626487732, "rewards/accuracy_reward/mean": 0.03629032149910927, "rewards/accuracy_reward/std": 0.1872003972530365, "step": 1420 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.798828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1717.0, "completions/mean_length": 1768.677734375, "completions/mean_terminated_length": 659.5242919921875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.48510710932832635, "frac_reward_zero_std": 0.78125, "grad_norm": 24.35814307083946, "learning_rate": 6.517763810375727e-07, "loss": 0.0059, "num_tokens": 1377694152.0, "reward": 0.064453125, "reward_std": 0.07564390450716019, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 1421 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.822265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1936.0, "completions/mean_length": 1831.720703125, "completions/mean_terminated_length": 831.1318969726562, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.48544849364171716, "frac_reward_zero_std": 0.78125, "grad_norm": 12.769176162152664, "learning_rate": 6.512540926525828e-07, "loss": 0.0276, "num_tokens": 1378708777.0, "reward": 0.01953125, "reward_std": 0.06233368441462517, "rewards/accuracy_reward/mean": 0.01953125, "rewards/accuracy_reward/std": 0.1385180652141571, "step": 1422 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1725.0, "completions/mean_length": 1858.42578125, "completions/mean_terminated_length": 770.868408203125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.485789877955108, "frac_reward_zero_std": 0.8125, "grad_norm": 33.136952907079994, "learning_rate": 6.507316605559281e-07, "loss": 0.0055, "num_tokens": 1379732259.0, "reward": 0.03515625, "reward_std": 0.06327171623706818, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 1423 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.826171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1961.0, "completions/mean_length": 1825.28515625, "completions/mean_terminated_length": 766.7640380859375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.4861312622684988, "frac_reward_zero_std": 0.75, "grad_norm": 21.755343588681697, "learning_rate": 6.502090854891051e-07, "loss": 0.0104, "num_tokens": 1380742181.0, "reward": 0.044921875, "reward_std": 0.08907270431518555, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 1424 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1880.0, "completions/mean_length": 1809.44140625, "completions/mean_terminated_length": 660.0227661132812, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.48647264658188955, "frac_reward_zero_std": 0.84375, "grad_norm": 9.755341671623205, "learning_rate": 6.496863681938138e-07, "loss": 0.0183, "num_tokens": 1381742087.0, "reward": 0.056640625, "reward_std": 0.055459219962358475, "rewards/accuracy_reward/mean": 0.058467742055654526, "rewards/accuracy_reward/std": 0.23486268520355225, "step": 1425 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.771484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.0, "completions/mean_length": 1738.595703125, "completions/mean_terminated_length": 694.0256958007812, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.48681403089528036, "frac_reward_zero_std": 0.84375, "grad_norm": 29.705652542540914, "learning_rate": 6.491635094119558e-07, "loss": 0.003, "num_tokens": 1382706872.0, "reward": 0.03125, "reward_std": 0.05880707502365112, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 1426 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.80078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2032.0, "completions/mean_length": 1774.232421875, "completions/mean_terminated_length": 673.7941284179688, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.48715541520867117, "frac_reward_zero_std": 0.8125, "grad_norm": 74.6742698270493, "learning_rate": 6.486405098856333e-07, "loss": 0.0113, "num_tokens": 1383695087.0, "reward": 0.017578125, "reward_std": 0.05452118441462517, "rewards/accuracy_reward/mean": 0.017578125, "rewards/accuracy_reward/std": 0.13154059648513794, "step": 1427 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2041.0, "completions/mean_length": 1796.181640625, "completions/mean_terminated_length": 704.96875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.487496799522062, "frac_reward_zero_std": 0.90625, "grad_norm": 20.94875629518616, "learning_rate": 6.481173703571487e-07, "loss": 0.0127, "num_tokens": 1384691820.0, "reward": 0.017578125, "reward_std": 0.03411140665411949, "rewards/accuracy_reward/mean": 0.017578125, "rewards/accuracy_reward/std": 0.13154059648513794, "step": 1428 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.84765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2013.0, "completions/mean_length": 1844.0859375, "completions/mean_terminated_length": 709.4871826171875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.48783818383545274, "frac_reward_zero_std": 0.84375, "grad_norm": 19.996064845265877, "learning_rate": 6.475940915690028e-07, "loss": 0.0073, "num_tokens": 1385717512.0, "reward": 0.0390625, "reward_std": 0.06162214279174805, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 1429 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.783203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2020.0, "completions/mean_length": 1752.017578125, "completions/mean_terminated_length": 682.7477416992188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.48817956814884356, "frac_reward_zero_std": 0.78125, "grad_norm": 22.784105088430692, "learning_rate": 6.470706742638942e-07, "loss": 0.0056, "num_tokens": 1386694689.0, "reward": 0.037109375, "reward_std": 0.08136672526597977, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 1430 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.744140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2008.0, "completions/mean_length": 1678.615234375, "completions/mean_terminated_length": 604.2977294921875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.48852095246223437, "frac_reward_zero_std": 0.65625, "grad_norm": 42.606477879428354, "learning_rate": 6.465471191847177e-07, "loss": 0.0419, "num_tokens": 1387629020.0, "reward": 0.080078125, "reward_std": 0.12713736295700073, "rewards/accuracy_reward/mean": 0.080078125, "rewards/accuracy_reward/std": 0.271679550409317, "step": 1431 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1972.0, "completions/mean_length": 1793.626953125, "completions/mean_terminated_length": 795.7019653320312, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.4888623367756252, "frac_reward_zero_std": 0.90625, "grad_norm": 32.05358850380085, "learning_rate": 6.460234270745645e-07, "loss": -0.0004, "num_tokens": 1388632333.0, "reward": 0.01953125, "reward_std": 0.03176242858171463, "rewards/accuracy_reward/mean": 0.01953125, "rewards/accuracy_reward/std": 0.1385180652141571, "step": 1432 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.69140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1776.0, "completions/mean_length": 1640.390625, "completions/mean_terminated_length": 727.1392822265625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.48920372108901594, "frac_reward_zero_std": 0.71875, "grad_norm": 17.148932696841907, "learning_rate": 6.454995986767193e-07, "loss": 0.018, "num_tokens": 1389546533.0, "reward": 0.060546875, "reward_std": 0.09793071448802948, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 1433 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.75, "completions/max_length": 2048.0, "completions/max_terminated_length": 1918.0, "completions/mean_length": 1721.84765625, "completions/mean_terminated_length": 743.390625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.48954510540240675, "frac_reward_zero_std": 0.75, "grad_norm": 17.835921685013446, "learning_rate": 6.44975634734661e-07, "loss": 0.0178, "num_tokens": 1390504583.0, "reward": 0.046875, "reward_std": 0.08698301762342453, "rewards/accuracy_reward/mean": 0.04838709533214569, "rewards/accuracy_reward/std": 0.21479946374893188, "step": 1434 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.732421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1912.0, "completions/mean_length": 1708.755859375, "completions/mean_terminated_length": 780.1678466796875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.48988648971579757, "frac_reward_zero_std": 0.6875, "grad_norm": 42.92018877470073, "learning_rate": 6.444515359920605e-07, "loss": 0.0059, "num_tokens": 1391455370.0, "reward": 0.078125, "reward_std": 0.11834745109081268, "rewards/accuracy_reward/mean": 0.078125, "rewards/accuracy_reward/std": 0.26863065361976624, "step": 1435 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1893.0, "completions/mean_length": 1779.349609375, "completions/mean_terminated_length": 774.8981323242188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.4902278740291884, "frac_reward_zero_std": 0.75, "grad_norm": 108.56796798241987, "learning_rate": 6.439273031927801e-07, "loss": 0.0044, "num_tokens": 1392463069.0, "reward": 0.064453125, "reward_std": 0.09138096868991852, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 1436 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2014.0, "completions/mean_length": 1674.857421875, "completions/mean_terminated_length": 683.3643188476562, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.49056925834257914, "frac_reward_zero_std": 0.90625, "grad_norm": 6.512244363605728, "learning_rate": 6.434029370808722e-07, "loss": 0.014, "num_tokens": 1393400100.0, "reward": 0.033203125, "reward_std": 0.044359706342220306, "rewards/accuracy_reward/mean": 0.03427419438958168, "rewards/accuracy_reward/std": 0.18211629986763, "step": 1437 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.76953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1987.0, "completions/mean_length": 1735.0703125, "completions/mean_terminated_length": 690.203369140625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.49091064265596995, "frac_reward_zero_std": 0.75, "grad_norm": 9.479888934581439, "learning_rate": 6.428784384005789e-07, "loss": 0.0111, "num_tokens": 1394365256.0, "reward": 0.05859375, "reward_std": 0.09381677210330963, "rewards/accuracy_reward/mean": 0.05859375, "rewards/accuracy_reward/std": 0.23509246110916138, "step": 1438 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1725.0, "completions/mean_length": 1743.16796875, "completions/mean_terminated_length": 702.5344848632812, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.49125202696936077, "frac_reward_zero_std": 0.71875, "grad_norm": 10.494916467439083, "learning_rate": 6.423538078963299e-07, "loss": 0.0197, "num_tokens": 1395340846.0, "reward": 0.1015625, "reward_std": 0.09666463732719421, "rewards/accuracy_reward/mean": 0.1015625, "rewards/accuracy_reward/std": 0.30236753821372986, "step": 1439 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2004.0, "completions/mean_length": 1679.01171875, "completions/mean_terminated_length": 754.013671875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.4915934112827516, "frac_reward_zero_std": 0.90625, "grad_norm": 36.713779456463264, "learning_rate": 6.418290463127423e-07, "loss": -0.0023, "num_tokens": 1396283268.0, "reward": 0.021484375, "reward_std": 0.036420635879039764, "rewards/accuracy_reward/mean": 0.02217741869390011, "rewards/accuracy_reward/std": 0.14740893244743347, "step": 1440 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.705078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1993.0, "completions/mean_length": 1657.201171875, "completions/mean_terminated_length": 723.6224975585938, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.49193479559614234, "frac_reward_zero_std": 0.625, "grad_norm": 108.78665438146012, "learning_rate": 6.413041543946192e-07, "loss": 0.0168, "num_tokens": 1397213515.0, "reward": 0.068359375, "reward_std": 0.1284080445766449, "rewards/accuracy_reward/mean": 0.068359375, "rewards/accuracy_reward/std": 0.25260838866233826, "step": 1441 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.73828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2044.0, "completions/mean_length": 1719.85546875, "completions/mean_terminated_length": 794.1940307617188, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.49227617990953315, "frac_reward_zero_std": 0.8125, "grad_norm": 55.40885216284095, "learning_rate": 6.407791328869488e-07, "loss": 0.0203, "num_tokens": 1398172673.0, "reward": 0.048828125, "reward_std": 0.07251107692718506, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 1442 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.724609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1960.0, "completions/mean_length": 1672.443359375, "completions/mean_terminated_length": 684.2765502929688, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.49261756422292396, "frac_reward_zero_std": 0.6875, "grad_norm": 3.2712755058134024, "learning_rate": 6.402539825349032e-07, "loss": 0.0227, "num_tokens": 1399101764.0, "reward": 0.099609375, "reward_std": 0.10931616276502609, "rewards/accuracy_reward/mean": 0.099609375, "rewards/accuracy_reward/std": 0.29977133870124817, "step": 1443 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.681640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2001.0, "completions/mean_length": 1612.3515625, "completions/mean_terminated_length": 679.5828247070312, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.4929589485363148, "frac_reward_zero_std": 0.8125, "grad_norm": 15.020980516220831, "learning_rate": 6.397287040838367e-07, "loss": 0.008, "num_tokens": 1399999704.0, "reward": 0.052734375, "reward_std": 0.06260748207569122, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 1444 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2007.0, "completions/mean_length": 1636.884765625, "completions/mean_terminated_length": 732.4312744140625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.49330033284970554, "frac_reward_zero_std": 0.75, "grad_norm": 15.35873829387781, "learning_rate": 6.392032982792865e-07, "loss": 0.0268, "num_tokens": 1400922413.0, "reward": 0.0859375, "reward_std": 0.09876786172389984, "rewards/accuracy_reward/mean": 0.0859375, "rewards/accuracy_reward/std": 0.28054583072662354, "step": 1445 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.759765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1997.0, "completions/mean_length": 1739.453125, "completions/mean_terminated_length": 763.6422729492188, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.49364171716309635, "frac_reward_zero_std": 0.71875, "grad_norm": 13.10752064545794, "learning_rate": 6.386777658669698e-07, "loss": 0.0165, "num_tokens": 1401887189.0, "reward": 0.072265625, "reward_std": 0.1141214370727539, "rewards/accuracy_reward/mean": 0.07459677755832672, "rewards/accuracy_reward/std": 0.263004869222641, "step": 1446 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.701171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1984.0, "completions/mean_length": 1684.115234375, "completions/mean_terminated_length": 830.2941284179688, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.49398310147648716, "frac_reward_zero_std": 0.75, "grad_norm": 9.225184352557372, "learning_rate": 6.38152107592784e-07, "loss": 0.0172, "num_tokens": 1402824896.0, "reward": 0.03515625, "reward_std": 0.08120594918727875, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 1447 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.822265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1610.0, "completions/mean_length": 1773.232421875, "completions/mean_terminated_length": 502.0549621582031, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.494324485789878, "frac_reward_zero_std": 0.90625, "grad_norm": 1.0264458006639359, "learning_rate": 6.376263242028048e-07, "loss": 0.0236, "num_tokens": 1403816023.0, "reward": 0.009765625, "reward_std": 0.028222277760505676, "rewards/accuracy_reward/mean": 0.009765625, "rewards/accuracy_reward/std": 0.09843364357948303, "step": 1448 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.744140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1717.0, "completions/mean_length": 1688.185546875, "completions/mean_terminated_length": 641.7022705078125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.49466587010326873, "frac_reward_zero_std": 0.75, "grad_norm": 2.203802785136163, "learning_rate": 6.371004164432853e-07, "loss": 0.031, "num_tokens": 1404751926.0, "reward": 0.033203125, "reward_std": 0.08082009106874466, "rewards/accuracy_reward/mean": 0.033203125, "rewards/accuracy_reward/std": 0.17934183776378632, "step": 1449 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.75390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2037.0, "completions/mean_length": 1728.521484375, "completions/mean_terminated_length": 749.8016357421875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.49500725441665955, "frac_reward_zero_std": 0.75, "grad_norm": 20.35667558109254, "learning_rate": 6.365743850606555e-07, "loss": 0.0044, "num_tokens": 1405715121.0, "reward": 0.05859375, "reward_std": 0.08302421122789383, "rewards/accuracy_reward/mean": 0.05859375, "rewards/accuracy_reward/std": 0.23509246110916138, "step": 1450 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.767578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2044.0, "completions/mean_length": 1726.2265625, "completions/mean_terminated_length": 663.5630493164062, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.49534863873005036, "frac_reward_zero_std": 0.78125, "grad_norm": 21.13413887002305, "learning_rate": 6.360482308015209e-07, "loss": 0.0173, "num_tokens": 1406685861.0, "reward": 0.03125, "reward_std": 0.07053204625844955, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 1451 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.818359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2008.0, "completions/mean_length": 1806.939453125, "completions/mean_terminated_length": 720.8709716796875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.4956900230434412, "frac_reward_zero_std": 0.8125, "grad_norm": 7.551419800574755, "learning_rate": 6.35521954412661e-07, "loss": -0.0036, "num_tokens": 1407691030.0, "reward": 0.041015625, "reward_std": 0.06183479726314545, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 1452 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.775390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2019.0, "completions/mean_length": 1770.41015625, "completions/mean_terminated_length": 812.5303955078125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.49603140735683193, "frac_reward_zero_std": 0.75, "grad_norm": 14.268487379277527, "learning_rate": 6.34995556641029e-07, "loss": 0.0111, "num_tokens": 1408676872.0, "reward": 0.080078125, "reward_std": 0.09841620922088623, "rewards/accuracy_reward/mean": 0.080078125, "rewards/accuracy_reward/std": 0.271679550409317, "step": 1453 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.826171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1492.0, "completions/mean_length": 1790.708984375, "completions/mean_terminated_length": 567.8539428710938, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.49637279167022275, "frac_reward_zero_std": 0.84375, "grad_norm": 1.394134018707034, "learning_rate": 6.344690382337503e-07, "loss": 0.0137, "num_tokens": 1409670659.0, "reward": 0.0234375, "reward_std": 0.053855981677770615, "rewards/accuracy_reward/mean": 0.0234375, "rewards/accuracy_reward/std": 0.15143637359142303, "step": 1454 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.767578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1736.0, "completions/mean_length": 1723.455078125, "completions/mean_terminated_length": 651.638671875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.49671417598361356, "frac_reward_zero_std": 0.75, "grad_norm": 16.69859558472023, "learning_rate": 6.339423999381216e-07, "loss": -0.0152, "num_tokens": 1410628156.0, "reward": 0.08203125, "reward_std": 0.09347065538167953, "rewards/accuracy_reward/mean": 0.08203125, "rewards/accuracy_reward/std": 0.2746807038784027, "step": 1455 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1952.0, "completions/mean_length": 1789.16796875, "completions/mean_terminated_length": 820.9444580078125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.4970555602970044, "frac_reward_zero_std": 0.8125, "grad_norm": 5.931127258141842, "learning_rate": 6.334156425016091e-07, "loss": 0.0168, "num_tokens": 1411620306.0, "reward": 0.044921875, "reward_std": 0.07481793314218521, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 1456 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.81640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1753.0, "completions/mean_length": 1799.6171875, "completions/mean_terminated_length": 695.1063842773438, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.49739694461039513, "frac_reward_zero_std": 0.75, "grad_norm": 9.026343701837094, "learning_rate": 6.328887666718493e-07, "loss": 0.0252, "num_tokens": 1412629998.0, "reward": 0.08984375, "reward_std": 0.10332095623016357, "rewards/accuracy_reward/mean": 0.08984375, "rewards/accuracy_reward/std": 0.2862374484539032, "step": 1457 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.81640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1888.0, "completions/mean_length": 1795.142578125, "completions/mean_terminated_length": 670.7340087890625, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.49773832892378594, "frac_reward_zero_std": 0.6875, "grad_norm": 5.745747347960597, "learning_rate": 6.323617731966456e-07, "loss": 0.0267, "num_tokens": 1413638855.0, "reward": 0.0390625, "reward_std": 0.09347065538167953, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 1458 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1855.0, "completions/mean_length": 1746.087890625, "completions/mean_terminated_length": 715.4224243164062, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.49807971323717676, "frac_reward_zero_std": 0.84375, "grad_norm": 30.858639232910065, "learning_rate": 6.318346628239691e-07, "loss": -0.001, "num_tokens": 1414616996.0, "reward": 0.041015625, "reward_std": 0.06502877175807953, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 1459 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.802734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2029.0, "completions/mean_length": 1785.61328125, "completions/mean_terminated_length": 717.8811645507812, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.49842109755056757, "frac_reward_zero_std": 0.78125, "grad_norm": 17.100907583510008, "learning_rate": 6.313074363019565e-07, "loss": 0.0109, "num_tokens": 1415603918.0, "reward": 0.029296875, "reward_std": 0.0714043527841568, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 1460 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.787109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1963.0, "completions/mean_length": 1756.369140625, "completions/mean_terminated_length": 678.1375732421875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.49876248186395833, "frac_reward_zero_std": 0.90625, "grad_norm": 1.96407771801503, "learning_rate": 6.307800943789093e-07, "loss": 0.0021, "num_tokens": 1416577371.0, "reward": 0.0234375, "reward_std": 0.03839729726314545, "rewards/accuracy_reward/mean": 0.0234375, "rewards/accuracy_reward/std": 0.15143637359142303, "step": 1461 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.76953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2048.0, "completions/mean_length": 1751.78515625, "completions/mean_terminated_length": 762.7288208007812, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.49910386617734914, "frac_reward_zero_std": 0.84375, "grad_norm": 6.895456310998595, "learning_rate": 6.302526378032931e-07, "loss": 0.0151, "num_tokens": 1417560893.0, "reward": 0.0703125, "reward_std": 0.06310540437698364, "rewards/accuracy_reward/mean": 0.0703125, "rewards/accuracy_reward/std": 0.25592297315597534, "step": 1462 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.806640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1963.0, "completions/mean_length": 1773.31640625, "completions/mean_terminated_length": 627.4141235351562, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.49944525049073996, "frac_reward_zero_std": 0.78125, "grad_norm": 3.6583850309125374, "learning_rate": 6.29725067323736e-07, "loss": 0.0258, "num_tokens": 1418546671.0, "reward": 0.0390625, "reward_std": 0.07588253915309906, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 1463 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1850.0, "completions/mean_length": 1803.166015625, "completions/mean_terminated_length": 623.5113525390625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.49978663480413077, "frac_reward_zero_std": 0.84375, "grad_norm": 2.2066377516828233, "learning_rate": 6.291973836890276e-07, "loss": 0.0207, "num_tokens": 1419548292.0, "reward": 0.01953125, "reward_std": 0.046875, "rewards/accuracy_reward/mean": 0.01953125, "rewards/accuracy_reward/std": 0.1385180652141571, "step": 1464 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1910.0, "completions/mean_length": 1623.7890625, "completions/mean_terminated_length": 539.6944580078125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.5001280191175216, "frac_reward_zero_std": 0.875, "grad_norm": 9.023797796566203, "learning_rate": 6.286695876481185e-07, "loss": 0.0093, "num_tokens": 1420460632.0, "reward": 0.044921875, "reward_std": 0.049345001578330994, "rewards/accuracy_reward/mean": 0.0463709682226181, "rewards/accuracy_reward/std": 0.21049949526786804, "step": 1465 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1935.0, "completions/mean_length": 1836.0078125, "completions/mean_terminated_length": 540.5, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5004694034309124, "frac_reward_zero_std": 0.875, "grad_norm": 1.3077878022213638, "learning_rate": 6.281416799501187e-07, "loss": 0.0046, "num_tokens": 1421476972.0, "reward": 0.021484375, "reward_std": 0.04604348540306091, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 1466 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.80859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1992.0, "completions/mean_length": 1769.732421875, "completions/mean_terminated_length": 594.69384765625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5008107877443031, "frac_reward_zero_std": 0.875, "grad_norm": 1.0235193741171722, "learning_rate": 6.276136613442964e-07, "loss": 0.0187, "num_tokens": 1422461539.0, "reward": 0.01953125, "reward_std": 0.04412011057138443, "rewards/accuracy_reward/mean": 0.02016128972172737, "rewards/accuracy_reward/std": 0.14069372415542603, "step": 1467 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.779296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1972.0, "completions/mean_length": 1747.501953125, "completions/mean_terminated_length": 686.4513549804688, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.5011521720576939, "frac_reward_zero_std": 0.8125, "grad_norm": 1.9995252362628586, "learning_rate": 6.270855325800775e-07, "loss": 0.0238, "num_tokens": 1423428404.0, "reward": 0.03515625, "reward_std": 0.06574726849794388, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 1468 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.794921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2022.0, "completions/mean_length": 1763.59375, "completions/mean_terminated_length": 661.1809692382812, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.5014935563710847, "frac_reward_zero_std": 0.78125, "grad_norm": 3.6900170609047698, "learning_rate": 6.265572944070444e-07, "loss": 0.025, "num_tokens": 1424400212.0, "reward": 0.04296875, "reward_std": 0.06964729726314545, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 1469 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.76953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1792.0, "completions/mean_length": 1720.244140625, "completions/mean_terminated_length": 625.8728637695312, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.5018349406844755, "frac_reward_zero_std": 0.65625, "grad_norm": 19.56069146226166, "learning_rate": 6.260289475749344e-07, "loss": 0.0393, "num_tokens": 1425351121.0, "reward": 0.0625, "reward_std": 0.12290477007627487, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.2422981858253479, "step": 1470 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1946.0, "completions/mean_length": 1803.080078125, "completions/mean_terminated_length": 623.0113525390625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5021763249978664, "frac_reward_zero_std": 0.78125, "grad_norm": 24.907489701801232, "learning_rate": 6.255004928336391e-07, "loss": 0.026, "num_tokens": 1426356042.0, "reward": 0.048828125, "reward_std": 0.08207826316356659, "rewards/accuracy_reward/mean": 0.05040322616696358, "rewards/accuracy_reward/std": 0.21899642050266266, "step": 1471 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1877.0, "completions/mean_length": 1815.953125, "completions/mean_terminated_length": 697.9091186523438, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.5025177093112572, "frac_reward_zero_std": 0.8125, "grad_norm": 24.951696580953566, "learning_rate": 6.249719309332036e-07, "loss": 0.0192, "num_tokens": 1427361810.0, "reward": 0.03515625, "reward_std": 0.06708566844463348, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 1472 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.77734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1978.0, "completions/mean_length": 1741.904296875, "completions/mean_terminated_length": 673.25439453125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.502859093624648, "frac_reward_zero_std": 0.75, "grad_norm": 28.344764999645086, "learning_rate": 6.244432626238245e-07, "loss": 0.019, "num_tokens": 1428327009.0, "reward": 0.052734375, "reward_std": 0.08890542387962341, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 1473 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.81640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1980.0, "completions/mean_length": 1814.224609375, "completions/mean_terminated_length": 775.223388671875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.5032004779380388, "frac_reward_zero_std": 0.875, "grad_norm": 4.979403647978972, "learning_rate": 6.239144886558501e-07, "loss": 0.0169, "num_tokens": 1429339860.0, "reward": 0.021484375, "reward_std": 0.04478531330823898, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 1474 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.83984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1968.0, "completions/mean_length": 1825.65625, "completions/mean_terminated_length": 660.3658447265625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.5035418622514295, "frac_reward_zero_std": 0.90625, "grad_norm": 20.798090517873316, "learning_rate": 6.23385609779778e-07, "loss": 0.006, "num_tokens": 1430362260.0, "reward": 0.021484375, "reward_std": 0.03462383896112442, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 1475 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1889.0, "completions/mean_length": 1734.498046875, "completions/mean_terminated_length": 664.2672119140625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5038832465648203, "frac_reward_zero_std": 0.84375, "grad_norm": 2.7091619444099373, "learning_rate": 6.228566267462555e-07, "loss": 0.0233, "num_tokens": 1431330339.0, "reward": 0.01953125, "reward_std": 0.05001020431518555, "rewards/accuracy_reward/mean": 0.01953125, "rewards/accuracy_reward/std": 0.1385180652141571, "step": 1476 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1641.0, "completions/mean_length": 1712.169921875, "completions/mean_terminated_length": 455.9166564941406, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.5042246308782111, "frac_reward_zero_std": 0.78125, "grad_norm": 4.908469758973546, "learning_rate": 6.22327540306077e-07, "loss": 0.0259, "num_tokens": 1432283674.0, "reward": 0.041015625, "reward_std": 0.07250870764255524, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 1477 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1959.0, "completions/mean_length": 1747.681640625, "completions/mean_terminated_length": 624.2684936523438, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.5045660151916019, "frac_reward_zero_std": 0.84375, "grad_norm": 45.99432201155855, "learning_rate": 6.217983512101838e-07, "loss": 0.0097, "num_tokens": 1433254551.0, "reward": 0.05859375, "reward_std": 0.07172322273254395, "rewards/accuracy_reward/mean": 0.05859375, "rewards/accuracy_reward/std": 0.23509246110916138, "step": 1478 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1636.0, "completions/mean_length": 1772.494140625, "completions/mean_terminated_length": 578.6354370117188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5049073995049927, "frac_reward_zero_std": 0.875, "grad_norm": 11.924320122773592, "learning_rate": 6.212690602096631e-07, "loss": 0.012, "num_tokens": 1434240052.0, "reward": 0.009765625, "reward_std": 0.03411140665411949, "rewards/accuracy_reward/mean": 0.009765625, "rewards/accuracy_reward/std": 0.09843364357948303, "step": 1479 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.84765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1824.0, "completions/mean_length": 1810.361328125, "completions/mean_terminated_length": 488.1153869628906, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5052487838183836, "frac_reward_zero_std": 0.78125, "grad_norm": 23.956286241026543, "learning_rate": 6.207396680557468e-07, "loss": 0.0097, "num_tokens": 1435245213.0, "reward": 0.05078125, "reward_std": 0.0908287987112999, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 1480 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.814453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1982.0, "completions/mean_length": 1790.21875, "completions/mean_terminated_length": 658.6947631835938, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.5055901681317744, "frac_reward_zero_std": 0.875, "grad_norm": 6.9881575822923265, "learning_rate": 6.202101754998101e-07, "loss": 0.0029, "num_tokens": 1436236381.0, "reward": 0.01953125, "reward_std": 0.04318207502365112, "rewards/accuracy_reward/mean": 0.01953125, "rewards/accuracy_reward/std": 0.1385180652141571, "step": 1481 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.85546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1959.0, "completions/mean_length": 1857.58984375, "completions/mean_terminated_length": 731.3513793945312, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.5059315524451652, "frac_reward_zero_std": 0.84375, "grad_norm": 10.524346440754835, "learning_rate": 6.196805832933709e-07, "loss": 0.0088, "num_tokens": 1437260347.0, "reward": 0.025390625, "reward_std": 0.05024883896112442, "rewards/accuracy_reward/mean": 0.025390625, "rewards/accuracy_reward/std": 0.15746226906776428, "step": 1482 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.775390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1958.0, "completions/mean_length": 1720.111328125, "completions/mean_terminated_length": 588.1825561523438, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.5062729367585559, "frac_reward_zero_std": 0.84375, "grad_norm": 13.445431303278404, "learning_rate": 6.191508921880886e-07, "loss": 0.0165, "num_tokens": 1438223188.0, "reward": 0.052734375, "reward_std": 0.07242262363433838, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 1483 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.810546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1996.0, "completions/mean_length": 1773.427734375, "completions/mean_terminated_length": 598.7113037109375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.5066143210719467, "frac_reward_zero_std": 0.8125, "grad_norm": 4.712775876981828, "learning_rate": 6.186211029357625e-07, "loss": 0.0081, "num_tokens": 1439205807.0, "reward": 0.021484375, "reward_std": 0.059305962175130844, "rewards/accuracy_reward/mean": 0.02217741869390011, "rewards/accuracy_reward/std": 0.14740893244743347, "step": 1484 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.80078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1979.0, "completions/mean_length": 1750.32421875, "completions/mean_terminated_length": 553.7843627929688, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.5069557053853375, "frac_reward_zero_std": 0.875, "grad_norm": 9.468026306558327, "learning_rate": 6.180912162883318e-07, "loss": 0.0103, "num_tokens": 1440181397.0, "reward": 0.013671875, "reward_std": 0.03889618441462517, "rewards/accuracy_reward/mean": 0.013671875, "rewards/accuracy_reward/std": 0.1162383034825325, "step": 1485 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.826171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1894.0, "completions/mean_length": 1803.833984375, "completions/mean_terminated_length": 643.8651733398438, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5072970896987283, "frac_reward_zero_std": 0.90625, "grad_norm": 7.155338504590691, "learning_rate": 6.175612329978737e-07, "loss": 0.0031, "num_tokens": 1441193168.0, "reward": 0.0078125, "reward_std": 0.02629890665411949, "rewards/accuracy_reward/mean": 0.0078125, "rewards/accuracy_reward/std": 0.08812850713729858, "step": 1486 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.84375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1974.0, "completions/mean_length": 1845.048828125, "completions/mean_terminated_length": 749.1124877929688, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5076384740121191, "frac_reward_zero_std": 0.90625, "grad_norm": 10.896039848112984, "learning_rate": 6.170311538166026e-07, "loss": 0.0109, "num_tokens": 1442212505.0, "reward": 0.0078125, "reward_std": 0.02629890665411949, "rewards/accuracy_reward/mean": 0.008064515888690948, "rewards/accuracy_reward/std": 0.0895301103591919, "step": 1487 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.830078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1791.0, "completions/mean_length": 1805.283203125, "completions/mean_terminated_length": 619.5977172851562, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.50797985832551, "frac_reward_zero_std": 0.84375, "grad_norm": 1.7314020114068187, "learning_rate": 6.165009794968687e-07, "loss": 0.02, "num_tokens": 1443215466.0, "reward": 0.029296875, "reward_std": 0.055033616721630096, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 1488 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.76953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1828.0, "completions/mean_length": 1690.1484375, "completions/mean_terminated_length": 495.28814697265625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5083212426389008, "frac_reward_zero_std": 0.78125, "grad_norm": 25.36217204271973, "learning_rate": 6.159707107911575e-07, "loss": 0.0358, "num_tokens": 1444159238.0, "reward": 0.072265625, "reward_std": 0.08885213732719421, "rewards/accuracy_reward/mean": 0.072265625, "rewards/accuracy_reward/std": 0.2591804563999176, "step": 1489 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.814453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1928.0, "completions/mean_length": 1778.5078125, "completions/mean_terminated_length": 595.5789794921875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5086626269522916, "frac_reward_zero_std": 0.6875, "grad_norm": 72.49000634756864, "learning_rate": 6.154403484520887e-07, "loss": 0.0332, "num_tokens": 1445145994.0, "reward": 0.0625, "reward_std": 0.10548296570777893, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.2422981858253479, "step": 1490 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2029.0, "completions/mean_length": 1791.0, "completions/mean_terminated_length": 617.7391357421875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5090040112656823, "frac_reward_zero_std": 0.875, "grad_norm": 1.7702183944556926, "learning_rate": 6.149098932324145e-07, "loss": 0.0085, "num_tokens": 1446135658.0, "reward": 0.037109375, "reward_std": 0.05440815910696983, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 1491 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.794921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1857.0, "completions/mean_length": 1746.35546875, "completions/mean_terminated_length": 577.1238403320312, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5093453955790731, "frac_reward_zero_std": 0.8125, "grad_norm": 9.448918934115191, "learning_rate": 6.143793458850188e-07, "loss": 0.019, "num_tokens": 1447101248.0, "reward": 0.05078125, "reward_std": 0.07419876754283905, "rewards/accuracy_reward/mean": 0.052419353276491165, "rewards/accuracy_reward/std": 0.22309619188308716, "step": 1492 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.779296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1778.0, "completions/mean_length": 1721.865234375, "completions/mean_terminated_length": 570.2920532226562, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.5096867798924639, "frac_reward_zero_std": 0.84375, "grad_norm": 11.820345441438167, "learning_rate": 6.13848707162917e-07, "loss": -0.0043, "num_tokens": 1448057403.0, "reward": 0.021484375, "reward_std": 0.05287160724401474, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 1493 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1696.0, "completions/mean_length": 1750.171875, "completions/mean_terminated_length": 636.0740966796875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5100281642058547, "frac_reward_zero_std": 0.875, "grad_norm": 0.21406744333176747, "learning_rate": 6.133179778192533e-07, "loss": 0.0213, "num_tokens": 1449020163.0, "reward": 0.017578125, "reward_std": 0.043135739862918854, "rewards/accuracy_reward/mean": 0.017578125, "rewards/accuracy_reward/std": 0.13154059648513794, "step": 1494 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.78515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2025.0, "completions/mean_length": 1745.009765625, "completions/mean_terminated_length": 637.7181396484375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5103695485192455, "frac_reward_zero_std": 0.71875, "grad_norm": 24.213039170025468, "learning_rate": 6.127871586073012e-07, "loss": 0.0192, "num_tokens": 1449994952.0, "reward": 0.048828125, "reward_std": 0.09331225603818893, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 1495 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.841796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1712.0, "completions/mean_length": 1810.13671875, "completions/mean_terminated_length": 544.4691162109375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5107109328326364, "frac_reward_zero_std": 0.90625, "grad_norm": 18.005339121716144, "learning_rate": 6.122562502804614e-07, "loss": -0.0027, "num_tokens": 1450989902.0, "reward": 0.03515625, "reward_std": 0.03944835811853409, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 1496 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.810546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1776.0, "completions/mean_length": 1784.84765625, "completions/mean_terminated_length": 658.9896850585938, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5110523171460272, "frac_reward_zero_std": 0.78125, "grad_norm": 23.81298234538911, "learning_rate": 6.117252535922611e-07, "loss": -0.0056, "num_tokens": 1451980448.0, "reward": 0.05859375, "reward_std": 0.08544550836086273, "rewards/accuracy_reward/mean": 0.05859375, "rewards/accuracy_reward/std": 0.23509246110916138, "step": 1497 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.78515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1994.0, "completions/mean_length": 1731.97265625, "completions/mean_terminated_length": 577.0363159179688, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.511393701459418, "frac_reward_zero_std": 0.71875, "grad_norm": 10.913167025446416, "learning_rate": 6.111941692963531e-07, "loss": 0.0008, "num_tokens": 1452951922.0, "reward": 0.046875, "reward_std": 0.10073335468769073, "rewards/accuracy_reward/mean": 0.046875, "rewards/accuracy_reward/std": 0.21157780289649963, "step": 1498 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.80078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1758.0, "completions/mean_length": 1758.140625, "completions/mean_terminated_length": 593.0196533203125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5117350857728087, "frac_reward_zero_std": 0.75, "grad_norm": 10.485520888574328, "learning_rate": 6.106629981465142e-07, "loss": 0.0067, "num_tokens": 1453929930.0, "reward": 0.033203125, "reward_std": 0.07686128467321396, "rewards/accuracy_reward/mean": 0.033203125, "rewards/accuracy_reward/std": 0.17934183776378632, "step": 1499 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1928.0, "completions/mean_length": 1707.71484375, "completions/mean_terminated_length": 546.0516967773438, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5120764700861995, "frac_reward_zero_std": 0.78125, "grad_norm": 11.499326610780647, "learning_rate": 6.101317408966451e-07, "loss": 0.0276, "num_tokens": 1454880344.0, "reward": 0.04296875, "reward_std": 0.08160631358623505, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 1500 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.814453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1624.0, "completions/mean_length": 1742.603515625, "completions/mean_terminated_length": 402.0736999511719, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5124178543995903, "frac_reward_zero_std": 0.84375, "grad_norm": 25.202970290657618, "learning_rate": 6.096003983007679e-07, "loss": 0.0021, "num_tokens": 1455846397.0, "reward": 0.0234375, "reward_std": 0.05259781330823898, "rewards/accuracy_reward/mean": 0.0234375, "rewards/accuracy_reward/std": 0.15143637359142303, "step": 1501 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.77734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2001.0, "completions/mean_length": 1725.25, "completions/mean_terminated_length": 598.4561157226562, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5127592387129811, "frac_reward_zero_std": 0.84375, "grad_norm": 5.899147768730763, "learning_rate": 6.090689711130263e-07, "loss": 0.013, "num_tokens": 1456804029.0, "reward": 0.033203125, "reward_std": 0.062062256038188934, "rewards/accuracy_reward/mean": 0.03427419438958168, "rewards/accuracy_reward/std": 0.18211629986763, "step": 1502 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.802734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1970.0, "completions/mean_length": 1783.8203125, "completions/mean_terminated_length": 708.7920532226562, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5131006230263719, "frac_reward_zero_std": 0.78125, "grad_norm": 6.990933728300156, "learning_rate": 6.085374600876842e-07, "loss": 0.024, "num_tokens": 1457791553.0, "reward": 0.0390625, "reward_std": 0.0792168527841568, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 1503 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1988.0, "completions/mean_length": 1796.939453125, "completions/mean_terminated_length": 650.7935180664062, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5134420073397628, "frac_reward_zero_std": 0.84375, "grad_norm": 4.335415013515333, "learning_rate": 6.080058659791241e-07, "loss": 0.0101, "num_tokens": 1458787106.0, "reward": 0.0234375, "reward_std": 0.05259781330823898, "rewards/accuracy_reward/mean": 0.024193547666072845, "rewards/accuracy_reward/std": 0.15380479395389557, "step": 1504 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.80859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2012.0, "completions/mean_length": 1761.2890625, "completions/mean_terminated_length": 550.0816040039062, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.5137833916531536, "frac_reward_zero_std": 0.84375, "grad_norm": 15.05040773508169, "learning_rate": 6.074741895418466e-07, "loss": -0.0007, "num_tokens": 1459769302.0, "reward": 0.025390625, "reward_std": 0.05859442427754402, "rewards/accuracy_reward/mean": 0.025390625, "rewards/accuracy_reward/std": 0.15746226906776428, "step": 1505 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.78125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2025.0, "completions/mean_length": 1758.458984375, "completions/mean_terminated_length": 724.3839721679688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5141247759665444, "frac_reward_zero_std": 0.84375, "grad_norm": 2.6576230895865556, "learning_rate": 6.069424315304693e-07, "loss": 0.0206, "num_tokens": 1460744817.0, "reward": 0.025390625, "reward_std": 0.04726085811853409, "rewards/accuracy_reward/mean": 0.025390625, "rewards/accuracy_reward/std": 0.15746226906776428, "step": 1506 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.806640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1967.0, "completions/mean_length": 1791.673828125, "completions/mean_terminated_length": 722.353515625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5144661602799352, "frac_reward_zero_std": 0.6875, "grad_norm": 6.69011436276421, "learning_rate": 6.064105926997251e-07, "loss": 0.042, "num_tokens": 1461730874.0, "reward": 0.03125, "reward_std": 0.09341736882925034, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 1507 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2023.0, "completions/mean_length": 1728.17578125, "completions/mean_terminated_length": 636.362060546875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5148075445933259, "frac_reward_zero_std": 0.78125, "grad_norm": 19.958448611887732, "learning_rate": 6.058786738044626e-07, "loss": 0.0099, "num_tokens": 1462694788.0, "reward": 0.064453125, "reward_std": 0.08452101051807404, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 1508 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.75, "completions/max_length": 2048.0, "completions/max_terminated_length": 1925.0, "completions/mean_length": 1713.025390625, "completions/mean_terminated_length": 708.1015625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.5151489289067167, "frac_reward_zero_std": 0.875, "grad_norm": 4.7108511684195555, "learning_rate": 6.053466755996427e-07, "loss": 0.0127, "num_tokens": 1463644881.0, "reward": 0.01953125, "reward_std": 0.04505910724401474, "rewards/accuracy_reward/mean": 0.01953125, "rewards/accuracy_reward/std": 0.1385180652141571, "step": 1509 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1797.0, "completions/mean_length": 1731.78515625, "completions/mean_terminated_length": 742.3386840820312, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5154903132201075, "frac_reward_zero_std": 0.78125, "grad_norm": 6.699401477668564, "learning_rate": 6.0481459884034e-07, "loss": 0.0338, "num_tokens": 1464602947.0, "reward": 0.05078125, "reward_std": 0.08214398473501205, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 1510 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.771484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2028.0, "completions/mean_length": 1737.35546875, "completions/mean_terminated_length": 689.0684204101562, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.5158316975334983, "frac_reward_zero_std": 0.90625, "grad_norm": 2.1213390779529826, "learning_rate": 6.042824442817399e-07, "loss": 0.022, "num_tokens": 1465566377.0, "reward": 0.0234375, "reward_std": 0.03839729726314545, "rewards/accuracy_reward/mean": 0.0234375, "rewards/accuracy_reward/std": 0.15143637359142303, "step": 1511 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.732421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1924.0, "completions/mean_length": 1666.064453125, "completions/mean_terminated_length": 620.6204223632812, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.5161730818468891, "frac_reward_zero_std": 0.8125, "grad_norm": 7.951404915232107, "learning_rate": 6.037502126791386e-07, "loss": 0.0044, "num_tokens": 1466493706.0, "reward": 0.029296875, "reward_std": 0.06403100490570068, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 1512 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.783203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1904.0, "completions/mean_length": 1731.083984375, "completions/mean_terminated_length": 586.189208984375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.51651446616028, "frac_reward_zero_std": 0.78125, "grad_norm": 7.2153715880723635, "learning_rate": 6.032179047879413e-07, "loss": 0.0289, "num_tokens": 1467460949.0, "reward": 0.072265625, "reward_std": 0.07603531330823898, "rewards/accuracy_reward/mean": 0.072265625, "rewards/accuracy_reward/std": 0.2591804563999176, "step": 1513 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.736328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2023.0, "completions/mean_length": 1671.61328125, "completions/mean_terminated_length": 620.5184936523438, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.5168558504736708, "frac_reward_zero_std": 0.78125, "grad_norm": 5.489564859370835, "learning_rate": 6.026855213636619e-07, "loss": 0.0211, "num_tokens": 1468388479.0, "reward": 0.0546875, "reward_std": 0.08471457660198212, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 1514 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1893.0, "completions/mean_length": 1728.66796875, "completions/mean_terminated_length": 809.3787841796875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5171972347870616, "frac_reward_zero_std": 0.6875, "grad_norm": 12.939105171186295, "learning_rate": 6.021530631619213e-07, "loss": 0.0116, "num_tokens": 1469348149.0, "reward": 0.06640625, "reward_std": 0.11206454783678055, "rewards/accuracy_reward/mean": 0.06854838877916336, "rewards/accuracy_reward/std": 0.25293973088264465, "step": 1515 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.775390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1934.0, "completions/mean_length": 1725.140625, "completions/mean_terminated_length": 610.5739135742188, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.5175386191004523, "frac_reward_zero_std": 0.6875, "grad_norm": 6.6046260277771465, "learning_rate": 6.016205309384466e-07, "loss": -0.0042, "num_tokens": 1470311725.0, "reward": 0.06640625, "reward_std": 0.12256665527820587, "rewards/accuracy_reward/mean": 0.06854838877916336, "rewards/accuracy_reward/std": 0.25293970108032227, "step": 1516 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.65234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1925.0, "completions/mean_length": 1565.03125, "completions/mean_terminated_length": 658.7864990234375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5178800034138431, "frac_reward_zero_std": 0.75, "grad_norm": 3.9368554451970024, "learning_rate": 6.010879254490695e-07, "loss": 0.0047, "num_tokens": 1471189677.0, "reward": 0.041015625, "reward_std": 0.09078246355056763, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 1517 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1884.0, "completions/mean_length": 1664.09375, "completions/mean_terminated_length": 683.0, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5182213877272339, "frac_reward_zero_std": 0.84375, "grad_norm": 1.8215671984381185, "learning_rate": 6.005552474497264e-07, "loss": 0.0111, "num_tokens": 1472119613.0, "reward": 0.03515625, "reward_std": 0.05914659798145294, "rewards/accuracy_reward/mean": 0.03629032149910927, "rewards/accuracy_reward/std": 0.1872003972530365, "step": 1518 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.73046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1863.0, "completions/mean_length": 1675.1640625, "completions/mean_terminated_length": 664.7246704101562, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5185627720406247, "frac_reward_zero_std": 0.71875, "grad_norm": 10.423043368875431, "learning_rate": 6.000224976964563e-07, "loss": 0.0334, "num_tokens": 1473059905.0, "reward": 0.048828125, "reward_std": 0.10188952833414078, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 1519 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1952.0, "completions/mean_length": 1701.271484375, "completions/mean_terminated_length": 703.1136474609375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.5189041563540155, "frac_reward_zero_std": 0.875, "grad_norm": 5.092466921206046, "learning_rate": 5.994896769453999e-07, "loss": 0.0015, "num_tokens": 1474008780.0, "reward": 0.03515625, "reward_std": 0.0479668527841568, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 1520 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.708984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1939.0, "completions/mean_length": 1649.09765625, "completions/mean_terminated_length": 677.275146484375, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.5192455406674064, "frac_reward_zero_std": 0.84375, "grad_norm": 2.9219695259100713, "learning_rate": 5.989567859527988e-07, "loss": 0.0188, "num_tokens": 1474930094.0, "reward": 0.01953125, "reward_std": 0.051493462175130844, "rewards/accuracy_reward/mean": 0.01953125, "rewards/accuracy_reward/std": 0.1385180652141571, "step": 1521 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.759765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1952.0, "completions/mean_length": 1717.294921875, "completions/mean_terminated_length": 671.406494140625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5195869249807972, "frac_reward_zero_std": 0.71875, "grad_norm": 3.668714818918523, "learning_rate": 5.984238254749946e-07, "loss": 0.0375, "num_tokens": 1475880965.0, "reward": 0.037109375, "reward_std": 0.09149399399757385, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 1522 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.744140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2048.0, "completions/mean_length": 1674.765625, "completions/mean_terminated_length": 589.2518920898438, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.519928309294188, "frac_reward_zero_std": 0.71875, "grad_norm": 3.191692323343434, "learning_rate": 5.978907962684267e-07, "loss": 0.0355, "num_tokens": 1476817229.0, "reward": 0.087890625, "reward_std": 0.09809703379869461, "rewards/accuracy_reward/mean": 0.087890625, "rewards/accuracy_reward/std": 0.2834126651287079, "step": 1523 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1983.0, "completions/mean_length": 1688.689453125, "completions/mean_terminated_length": 695.3014526367188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5202696936075787, "frac_reward_zero_std": 0.84375, "grad_norm": 16.394918768435236, "learning_rate": 5.973576990896331e-07, "loss": 0.0098, "num_tokens": 1477751838.0, "reward": 0.0234375, "reward_std": 0.05667105317115784, "rewards/accuracy_reward/mean": 0.0234375, "rewards/accuracy_reward/std": 0.15143637359142303, "step": 1524 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1942.0, "completions/mean_length": 1738.240234375, "completions/mean_terminated_length": 579.50927734375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5206110779209695, "frac_reward_zero_std": 0.78125, "grad_norm": 20.843726607976432, "learning_rate": 5.968245346952473e-07, "loss": 0.0041, "num_tokens": 1478720553.0, "reward": 0.037109375, "reward_std": 0.07355976104736328, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 1525 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.76953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1924.0, "completions/mean_length": 1711.24609375, "completions/mean_terminated_length": 587.2626953125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5209524622343603, "frac_reward_zero_std": 0.71875, "grad_norm": 17.61273419512016, "learning_rate": 5.962913038419988e-07, "loss": -0.001, "num_tokens": 1479676855.0, "reward": 0.0625, "reward_std": 0.10161572694778442, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.2422981858253479, "step": 1526 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.673828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1793.0, "completions/mean_length": 1603.716796875, "completions/mean_terminated_length": 685.88623046875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5212938465477511, "frac_reward_zero_std": 0.8125, "grad_norm": 5.049159518715146, "learning_rate": 5.957580072867113e-07, "loss": 0.0215, "num_tokens": 1480572582.0, "reward": 0.052734375, "reward_std": 0.08298446238040924, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 1527 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.685546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1824.0, "completions/mean_length": 1620.80859375, "completions/mean_terminated_length": 689.478271484375, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.5216352308611419, "frac_reward_zero_std": 0.8125, "grad_norm": 5.5573006106581655, "learning_rate": 5.952246457863019e-07, "loss": 0.0206, "num_tokens": 1481475028.0, "reward": 0.041015625, "reward_std": 0.07285481691360474, "rewards/accuracy_reward/mean": 0.04233871027827263, "rewards/accuracy_reward/std": 0.2015640139579773, "step": 1528 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.728515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1820.0, "completions/mean_length": 1657.556640625, "completions/mean_terminated_length": 609.8201293945312, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.5219766151745328, "frac_reward_zero_std": 0.71875, "grad_norm": 8.563085377122176, "learning_rate": 5.946912200977794e-07, "loss": 0.0133, "num_tokens": 1482406065.0, "reward": 0.060546875, "reward_std": 0.10453042387962341, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 1529 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.732421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1976.0, "completions/mean_length": 1669.744140625, "completions/mean_terminated_length": 634.3722534179688, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.5223179994879236, "frac_reward_zero_std": 0.75, "grad_norm": 6.101719864674231, "learning_rate": 5.941577309782441e-07, "loss": 0.0118, "num_tokens": 1483341774.0, "reward": 0.0546875, "reward_std": 0.08493966609239578, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 1530 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.732421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1910.0, "completions/mean_length": 1682.529296875, "completions/mean_terminated_length": 682.1532592773438, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5226593838013144, "frac_reward_zero_std": 0.65625, "grad_norm": 1.2187278258013974, "learning_rate": 5.936241791848863e-07, "loss": 0.0497, "num_tokens": 1484276509.0, "reward": 0.06640625, "reward_std": 0.12846125662326813, "rewards/accuracy_reward/mean": 0.06640625, "rewards/accuracy_reward/std": 0.2492343932390213, "step": 1531 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1952.0, "completions/mean_length": 1661.78125, "completions/mean_terminated_length": 635.5428466796875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5230007681147051, "frac_reward_zero_std": 0.78125, "grad_norm": 11.92125611976677, "learning_rate": 5.930905654749848e-07, "loss": 0.0103, "num_tokens": 1485204893.0, "reward": 0.0390625, "reward_std": 0.08109388500452042, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 1532 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2014.0, "completions/mean_length": 1701.076171875, "completions/mean_terminated_length": 741.933837890625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5233421524280959, "frac_reward_zero_std": 0.6875, "grad_norm": 6.014843141083682, "learning_rate": 5.925568906059073e-07, "loss": 0.0373, "num_tokens": 1486150404.0, "reward": 0.0703125, "reward_std": 0.12145408242940903, "rewards/accuracy_reward/mean": 0.0703125, "rewards/accuracy_reward/std": 0.25592297315597534, "step": 1533 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.708984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2040.0, "completions/mean_length": 1658.205078125, "completions/mean_terminated_length": 708.9395751953125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.5236835367414867, "frac_reward_zero_std": 0.625, "grad_norm": 10.044092412160285, "learning_rate": 5.920231553351073e-07, "loss": 0.0304, "num_tokens": 1487083309.0, "reward": 0.056640625, "reward_std": 0.12636469304561615, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 1534 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.74609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1968.0, "completions/mean_length": 1681.025390625, "completions/mean_terminated_length": 603.0076904296875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5240249210548775, "frac_reward_zero_std": 0.6875, "grad_norm": 21.288342537773524, "learning_rate": 5.914893604201244e-07, "loss": 0.0406, "num_tokens": 1488033226.0, "reward": 0.07421875, "reward_std": 0.11917342245578766, "rewards/accuracy_reward/mean": 0.07421875, "rewards/accuracy_reward/std": 0.2623828947544098, "step": 1535 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.728515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1866.0, "completions/mean_length": 1669.74609375, "completions/mean_terminated_length": 654.7194213867188, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5243663053682683, "frac_reward_zero_std": 0.75, "grad_norm": 2.162597788905748, "learning_rate": 5.909555066185829e-07, "loss": 0.0094, "num_tokens": 1488966712.0, "reward": 0.021484375, "reward_std": 0.07108421623706818, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 1536 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.666015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.0, "completions/mean_length": 1591.837890625, "completions/mean_terminated_length": 682.1812744140625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5247076896816592, "frac_reward_zero_std": 0.78125, "grad_norm": 6.067570829059677, "learning_rate": 5.904215946881907e-07, "loss": 0.012, "num_tokens": 1489861173.0, "reward": 0.056640625, "reward_std": 0.08070706576108932, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 1537 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.724609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1989.0, "completions/mean_length": 1672.943359375, "completions/mean_terminated_length": 686.0921630859375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.52504907399505, "frac_reward_zero_std": 0.71875, "grad_norm": 6.2500102276916465, "learning_rate": 5.898876253867379e-07, "loss": 0.014, "num_tokens": 1490793224.0, "reward": 0.080078125, "reward_std": 0.12055579572916031, "rewards/accuracy_reward/mean": 0.080078125, "rewards/accuracy_reward/std": 0.271679550409317, "step": 1538 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1864.0, "completions/mean_length": 1692.986328125, "completions/mean_terminated_length": 749.664306640625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5253904583084408, "frac_reward_zero_std": 0.71875, "grad_norm": 6.400671788590852, "learning_rate": 5.893535994720965e-07, "loss": 0.0097, "num_tokens": 1491736145.0, "reward": 0.09375, "reward_std": 0.10563018918037415, "rewards/accuracy_reward/mean": 0.09375, "rewards/accuracy_reward/std": 0.29176566004753113, "step": 1539 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.73046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1817.0, "completions/mean_length": 1678.689453125, "completions/mean_terminated_length": 677.8043823242188, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.5257318426218315, "frac_reward_zero_std": 0.71875, "grad_norm": 9.811172391020396, "learning_rate": 5.888195177022185e-07, "loss": 0.0273, "num_tokens": 1492673282.0, "reward": 0.0546875, "reward_std": 0.10139063745737076, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 1540 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.72265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1945.0, "completions/mean_length": 1673.94140625, "completions/mean_terminated_length": 699.2816772460938, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.5260732269352223, "frac_reward_zero_std": 0.65625, "grad_norm": 23.88244549271511, "learning_rate": 5.882853808351354e-07, "loss": 0.0326, "num_tokens": 1493603300.0, "reward": 0.0703125, "reward_std": 0.12538030743598938, "rewards/accuracy_reward/mean": 0.07258064299821854, "rewards/accuracy_reward/std": 0.25970885157585144, "step": 1541 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.748046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1815.0, "completions/mean_length": 1701.55859375, "completions/mean_terminated_length": 672.9767456054688, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.5264146112486131, "frac_reward_zero_std": 0.8125, "grad_norm": 10.203258176141997, "learning_rate": 5.877511896289566e-07, "loss": 0.02, "num_tokens": 1494548610.0, "reward": 0.025390625, "reward_std": 0.06354551017284393, "rewards/accuracy_reward/mean": 0.025390625, "rewards/accuracy_reward/std": 0.15746226906776428, "step": 1542 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.806640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2000.0, "completions/mean_length": 1795.78515625, "completions/mean_terminated_length": 744.0504760742188, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.5267559955620039, "frac_reward_zero_std": 0.75, "grad_norm": 28.33020389871299, "learning_rate": 5.872169448418688e-07, "loss": 0.0267, "num_tokens": 1495555540.0, "reward": 0.05859375, "reward_std": 0.09056951105594635, "rewards/accuracy_reward/mean": 0.05859375, "rewards/accuracy_reward/std": 0.23509246110916138, "step": 1543 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.740234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1801.0, "completions/mean_length": 1699.875, "completions/mean_terminated_length": 707.8496704101562, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5270973798753947, "frac_reward_zero_std": 0.78125, "grad_norm": 67.1877398251019, "learning_rate": 5.866826472321351e-07, "loss": 0.0186, "num_tokens": 1496504276.0, "reward": 0.037109375, "reward_std": 0.07768725603818893, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 1544 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.736328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1968.0, "completions/mean_length": 1681.78515625, "completions/mean_terminated_length": 659.0963134765625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5274387641887855, "frac_reward_zero_std": 0.78125, "grad_norm": 6.846714051394489, "learning_rate": 5.861482975580928e-07, "loss": 0.0124, "num_tokens": 1497450102.0, "reward": 0.021484375, "reward_std": 0.06425705552101135, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 1545 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.72265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1918.0, "completions/mean_length": 1690.87890625, "completions/mean_terminated_length": 760.3521118164062, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.5277801485021764, "frac_reward_zero_std": 0.75, "grad_norm": 7.06535709332671, "learning_rate": 5.856138965781538e-07, "loss": 0.0249, "num_tokens": 1498392120.0, "reward": 0.05859375, "reward_std": 0.09586012363433838, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.24231401085853577, "step": 1546 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.763671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2032.0, "completions/mean_length": 1720.404296875, "completions/mean_terminated_length": 661.8098754882812, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.5281215328155672, "frac_reward_zero_std": 0.875, "grad_norm": 6.72553105609815, "learning_rate": 5.850794450508026e-07, "loss": 0.0086, "num_tokens": 1499353399.0, "reward": 0.03125, "reward_std": 0.055114150047302246, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 1547 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.77734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1865.0, "completions/mean_length": 1741.125, "completions/mean_terminated_length": 669.75439453125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.5284629171289579, "frac_reward_zero_std": 0.84375, "grad_norm": 8.088981743521808, "learning_rate": 5.845449437345952e-07, "loss": 0.0142, "num_tokens": 1500327399.0, "reward": 0.025390625, "reward_std": 0.05298367142677307, "rewards/accuracy_reward/mean": 0.025390625, "rewards/accuracy_reward/std": 0.15746226906776428, "step": 1548 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.779296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1814.0, "completions/mean_length": 1718.646484375, "completions/mean_terminated_length": 555.7079467773438, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5288043014423487, "frac_reward_zero_std": 0.71875, "grad_norm": 21.943762028457954, "learning_rate": 5.840103933881584e-07, "loss": 0.0187, "num_tokens": 1501284866.0, "reward": 0.06640625, "reward_std": 0.1165834367275238, "rewards/accuracy_reward/mean": 0.06640625, "rewards/accuracy_reward/std": 0.2492343932390213, "step": 1549 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.80078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1912.0, "completions/mean_length": 1776.9453125, "completions/mean_terminated_length": 687.4118041992188, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5291456857557395, "frac_reward_zero_std": 0.8125, "grad_norm": 6.719228970002369, "learning_rate": 5.834757947701889e-07, "loss": 0.0175, "num_tokens": 1502278950.0, "reward": 0.04296875, "reward_std": 0.051659777760505676, "rewards/accuracy_reward/mean": 0.04435483738780022, "rewards/accuracy_reward/std": 0.2060900777578354, "step": 1550 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.814453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1920.0, "completions/mean_length": 1794.490234375, "completions/mean_terminated_length": 682.252685546875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5294870700691303, "frac_reward_zero_std": 0.84375, "grad_norm": 12.281057050240783, "learning_rate": 5.829411486394516e-07, "loss": 0.0119, "num_tokens": 1503280385.0, "reward": 0.03125, "reward_std": 0.05789502337574959, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 1551 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.810546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2004.0, "completions/mean_length": 1776.78515625, "completions/mean_terminated_length": 616.4329833984375, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.5298284543825211, "frac_reward_zero_std": 0.75, "grad_norm": 7.894476504088849, "learning_rate": 5.824064557547785e-07, "loss": 0.0144, "num_tokens": 1504259875.0, "reward": 0.064453125, "reward_std": 0.08505964279174805, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 1552 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.75, "completions/max_length": 2048.0, "completions/max_terminated_length": 2038.0, "completions/mean_length": 1682.591796875, "completions/mean_terminated_length": 586.3671875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5301698386959119, "frac_reward_zero_std": 0.8125, "grad_norm": 6.709904486595488, "learning_rate": 5.81871716875069e-07, "loss": 0.0142, "num_tokens": 1505205138.0, "reward": 0.03125, "reward_std": 0.06079617142677307, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 1553 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.77734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1938.0, "completions/mean_length": 1749.833984375, "completions/mean_terminated_length": 708.868408203125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5305112230093028, "frac_reward_zero_std": 0.90625, "grad_norm": 11.595723159705539, "learning_rate": 5.813369327592867e-07, "loss": 0.0075, "num_tokens": 1506182157.0, "reward": 0.02734375, "reward_std": 0.03779878467321396, "rewards/accuracy_reward/mean": 0.02822580561041832, "rewards/accuracy_reward/std": 0.1657845675945282, "step": 1554 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1790.0, "completions/mean_length": 1786.193359375, "completions/mean_terminated_length": 651.6979370117188, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.5308526073226936, "frac_reward_zero_std": 0.78125, "grad_norm": 7.7419450565362995, "learning_rate": 5.808021041664599e-07, "loss": 0.0306, "num_tokens": 1507173136.0, "reward": 0.03125, "reward_std": 0.07300759106874466, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 1555 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.72265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1992.0, "completions/mean_length": 1654.302734375, "completions/mean_terminated_length": 628.4718017578125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5311939916360843, "frac_reward_zero_std": 0.625, "grad_norm": 30.72497500795339, "learning_rate": 5.802672318556802e-07, "loss": 0.0153, "num_tokens": 1508093307.0, "reward": 0.0703125, "reward_std": 0.11740703880786896, "rewards/accuracy_reward/mean": 0.0703125, "rewards/accuracy_reward/std": 0.25592297315597534, "step": 1556 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.806640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1685.0, "completions/mean_length": 1750.765625, "completions/mean_terminated_length": 510.7878723144531, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5315353759494751, "frac_reward_zero_std": 0.75, "grad_norm": 6.90278174268782, "learning_rate": 5.797323165861007e-07, "loss": 0.0075, "num_tokens": 1509070195.0, "reward": 0.041015625, "reward_std": 0.07729348540306091, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 1557 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.798828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1989.0, "completions/mean_length": 1750.853515625, "completions/mean_terminated_length": 570.92236328125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5318767602628659, "frac_reward_zero_std": 0.75, "grad_norm": 18.750295588036277, "learning_rate": 5.791973591169359e-07, "loss": 0.0242, "num_tokens": 1510047112.0, "reward": 0.068359375, "reward_std": 0.09903506934642792, "rewards/accuracy_reward/mean": 0.068359375, "rewards/accuracy_reward/std": 0.25260838866233826, "step": 1558 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.826171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2025.0, "completions/mean_length": 1814.37890625, "completions/mean_terminated_length": 704.0224609375, "completions/min_length": 7.0, "completions/min_terminated_length": 7.0, "epoch": 0.5322181445762567, "frac_reward_zero_std": 0.75, "grad_norm": 4.311757674099231, "learning_rate": 5.786623602074602e-07, "loss": 0.0366, "num_tokens": 1511045946.0, "reward": 0.046875, "reward_std": 0.09045085310935974, "rewards/accuracy_reward/mean": 0.046875, "rewards/accuracy_reward/std": 0.21157780289649963, "step": 1559 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.79296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1903.0, "completions/mean_length": 1751.97265625, "completions/mean_terminated_length": 618.132080078125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.5325595288896475, "frac_reward_zero_std": 0.84375, "grad_norm": 10.408238486982613, "learning_rate": 5.781273206170065e-07, "loss": 0.0172, "num_tokens": 1512025756.0, "reward": 0.044921875, "reward_std": 0.06118203327059746, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 1560 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.78515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1794.0, "completions/mean_length": 1757.056640625, "completions/mean_terminated_length": 693.7908935546875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5329009132030383, "frac_reward_zero_std": 0.8125, "grad_norm": 13.732166503274463, "learning_rate": 5.775922411049657e-07, "loss": -0.0018, "num_tokens": 1513014281.0, "reward": 0.0234375, "reward_std": 0.05754890665411949, "rewards/accuracy_reward/mean": 0.0234375, "rewards/accuracy_reward/std": 0.15143637359142303, "step": 1561 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.70703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1824.0, "completions/mean_length": 1625.11328125, "completions/mean_terminated_length": 604.5466918945312, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.5332422975164292, "frac_reward_zero_std": 0.75, "grad_norm": 9.69746032400128, "learning_rate": 5.770571224307855e-07, "loss": 0.0396, "num_tokens": 1513919123.0, "reward": 0.0859375, "reward_std": 0.11326298117637634, "rewards/accuracy_reward/mean": 0.08870967477560043, "rewards/accuracy_reward/std": 0.284611314535141, "step": 1562 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.787109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1788.0, "completions/mean_length": 1758.107421875, "completions/mean_terminated_length": 686.302734375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.53358368182982, "frac_reward_zero_std": 0.78125, "grad_norm": 11.219515853370563, "learning_rate": 5.765219653539687e-07, "loss": 0.0158, "num_tokens": 1514894714.0, "reward": 0.021484375, "reward_std": 0.06425705552101135, "rewards/accuracy_reward/mean": 0.02217741869390011, "rewards/accuracy_reward/std": 0.14740893244743347, "step": 1563 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.79296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1600.0, "completions/mean_length": 1732.203125, "completions/mean_terminated_length": 522.6415405273438, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5339250661432107, "frac_reward_zero_std": 0.875, "grad_norm": 3.81842005871031, "learning_rate": 5.759867706340731e-07, "loss": 0.0181, "num_tokens": 1515865394.0, "reward": 0.029296875, "reward_std": 0.04726085811853409, "rewards/accuracy_reward/mean": 0.030241934582591057, "rewards/accuracy_reward/std": 0.1714252382516861, "step": 1564 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.837890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1835.0, "completions/mean_length": 1809.412109375, "completions/mean_terminated_length": 576.2288818359375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5342664504566015, "frac_reward_zero_std": 0.8125, "grad_norm": 11.857555351983311, "learning_rate": 5.754515390307095e-07, "loss": 0.0131, "num_tokens": 1516872469.0, "reward": 0.033203125, "reward_std": 0.06508206576108932, "rewards/accuracy_reward/mean": 0.033203125, "rewards/accuracy_reward/std": 0.17934183776378632, "step": 1565 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1838.0, "completions/mean_length": 1766.927734375, "completions/mean_terminated_length": 548.9479370117188, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5346078347699923, "frac_reward_zero_std": 0.875, "grad_norm": 6.325629083931258, "learning_rate": 5.749162713035415e-07, "loss": -0.003, "num_tokens": 1517853936.0, "reward": 0.0234375, "reward_std": 0.04670868441462517, "rewards/accuracy_reward/mean": 0.0234375, "rewards/accuracy_reward/std": 0.15143637359142303, "step": 1566 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.837890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1959.0, "completions/mean_length": 1824.177734375, "completions/mean_terminated_length": 667.9879150390625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5349492190833831, "frac_reward_zero_std": 0.71875, "grad_norm": 46.115858137580624, "learning_rate": 5.743809682122836e-07, "loss": 0.0153, "num_tokens": 1518869947.0, "reward": 0.03515625, "reward_std": 0.08709508180618286, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 1567 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.77734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1920.0, "completions/mean_length": 1730.818359375, "completions/mean_terminated_length": 623.4649047851562, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5352906033967739, "frac_reward_zero_std": 0.65625, "grad_norm": 22.45428717285173, "learning_rate": 5.738456305167007e-07, "loss": 0.0208, "num_tokens": 1519831118.0, "reward": 0.05078125, "reward_std": 0.1049705445766449, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 1568 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.75390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1829.0, "completions/mean_length": 1682.189453125, "completions/mean_terminated_length": 561.5317993164062, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5356319877101647, "frac_reward_zero_std": 0.78125, "grad_norm": 14.612831349622036, "learning_rate": 5.733102589766068e-07, "loss": -0.0029, "num_tokens": 1520761647.0, "reward": 0.0390625, "reward_std": 0.07795868813991547, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 1569 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.75390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1759.0, "completions/mean_length": 1704.345703125, "completions/mean_terminated_length": 651.5635375976562, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5359733720235555, "frac_reward_zero_std": 0.6875, "grad_norm": 42.354342888795586, "learning_rate": 5.727748543518637e-07, "loss": 0.0238, "num_tokens": 1521711024.0, "reward": 0.052734375, "reward_std": 0.10834437608718872, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 1570 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.83203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1966.0, "completions/mean_length": 1802.828125, "completions/mean_terminated_length": 588.3720703125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5363147563369464, "frac_reward_zero_std": 0.6875, "grad_norm": 10.871412453464968, "learning_rate": 5.722394174023805e-07, "loss": 0.0168, "num_tokens": 1522713000.0, "reward": 0.052734375, "reward_std": 0.10854348540306091, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 1571 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.779296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1958.0, "completions/mean_length": 1722.22265625, "completions/mean_terminated_length": 571.9114990234375, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.5366561406503371, "frac_reward_zero_std": 0.6875, "grad_norm": 98.84162473141893, "learning_rate": 5.717039488881118e-07, "loss": 0.0079, "num_tokens": 1523670970.0, "reward": 0.052734375, "reward_std": 0.10453139245510101, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 1572 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1799.0, "completions/mean_length": 1753.447265625, "completions/mean_terminated_length": 597.8942260742188, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.5369975249637279, "frac_reward_zero_std": 0.75, "grad_norm": 29.69815838452254, "learning_rate": 5.711684495690573e-07, "loss": 0.0119, "num_tokens": 1524638671.0, "reward": 0.0390625, "reward_std": 0.08263043314218521, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 1573 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.78515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1759.0, "completions/mean_length": 1745.953125, "completions/mean_terminated_length": 642.1090698242188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5373389092771187, "frac_reward_zero_std": 0.78125, "grad_norm": 84.9726769521578, "learning_rate": 5.706329202052605e-07, "loss": 0.006, "num_tokens": 1525610199.0, "reward": 0.03515625, "reward_std": 0.07531681656837463, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 1574 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.69140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2017.0, "completions/mean_length": 1609.865234375, "completions/mean_terminated_length": 628.2215576171875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5376802935905095, "frac_reward_zero_std": 0.78125, "grad_norm": 3.304391370281301, "learning_rate": 5.700973615568072e-07, "loss": 0.0342, "num_tokens": 1526509426.0, "reward": 0.029296875, "reward_std": 0.06587383896112442, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 1575 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.775390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2007.0, "completions/mean_length": 1726.68359375, "completions/mean_terminated_length": 617.8869018554688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5380216779039003, "frac_reward_zero_std": 0.8125, "grad_norm": 17.33273339545191, "learning_rate": 5.695617743838252e-07, "loss": 0.0017, "num_tokens": 1527475680.0, "reward": 0.017578125, "reward_std": 0.05452118441462517, "rewards/accuracy_reward/mean": 0.017578125, "rewards/accuracy_reward/std": 0.13154059648513794, "step": 1576 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.826171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2012.0, "completions/mean_length": 1799.552734375, "completions/mean_terminated_length": 618.7303466796875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5383630622172911, "frac_reward_zero_std": 0.78125, "grad_norm": 21.48197827741597, "learning_rate": 5.690261594464824e-07, "loss": 0.0108, "num_tokens": 1528467179.0, "reward": 0.044921875, "reward_std": 0.07729348540306091, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 1577 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.783203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1996.0, "completions/mean_length": 1747.06640625, "completions/mean_terminated_length": 659.909912109375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.538704446530682, "frac_reward_zero_std": 0.9375, "grad_norm": 39.89398079801477, "learning_rate": 5.68490517504986e-07, "loss": -0.0026, "num_tokens": 1529439005.0, "reward": 0.029296875, "reward_std": 0.03214829042553902, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 1578 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1781.0, "completions/mean_length": 1632.5390625, "completions/mean_terminated_length": 648.5526123046875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5390458308440728, "frac_reward_zero_std": 0.84375, "grad_norm": 8.495759615266493, "learning_rate": 5.67954849319582e-07, "loss": 0.0231, "num_tokens": 1530355585.0, "reward": 0.029296875, "reward_std": 0.05298367142677307, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 1579 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.775390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1877.0, "completions/mean_length": 1730.712890625, "completions/mean_terminated_length": 635.382568359375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.5393872151574635, "frac_reward_zero_std": 0.84375, "grad_norm": 8.934966614858316, "learning_rate": 5.674191556505533e-07, "loss": 0.0126, "num_tokens": 1531329934.0, "reward": 0.01953125, "reward_std": 0.046875, "rewards/accuracy_reward/mean": 0.01953125, "rewards/accuracy_reward/std": 0.1385180652141571, "step": 1580 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.791015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1861.0, "completions/mean_length": 1763.912109375, "completions/mean_terminated_length": 689.168212890625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.5397285994708543, "frac_reward_zero_std": 0.8125, "grad_norm": 7.372383976573982, "learning_rate": 5.668834372582195e-07, "loss": 0.0043, "num_tokens": 1532318017.0, "reward": 0.0390625, "reward_std": 0.06888246536254883, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 1581 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.806640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1944.0, "completions/mean_length": 1747.025390625, "completions/mean_terminated_length": 492.0303039550781, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5400699837842451, "frac_reward_zero_std": 0.625, "grad_norm": 27.415553284280957, "learning_rate": 5.663476949029342e-07, "loss": -0.0069, "num_tokens": 1533296318.0, "reward": 0.06640625, "reward_std": 0.12230625748634338, "rewards/accuracy_reward/mean": 0.06640625, "rewards/accuracy_reward/std": 0.2492343932390213, "step": 1582 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1562.0, "completions/mean_length": 1764.556640625, "completions/mean_terminated_length": 596.7699584960938, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5404113680976359, "frac_reward_zero_std": 0.875, "grad_norm": 16.497228299495884, "learning_rate": 5.65811929345086e-07, "loss": 0.0076, "num_tokens": 1534266027.0, "reward": 0.02734375, "reward_std": 0.04907120391726494, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 1583 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2016.0, "completions/mean_length": 1667.283203125, "completions/mean_terminated_length": 571.2803344726562, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.5407527524110267, "frac_reward_zero_std": 0.75, "grad_norm": 11.331687761976324, "learning_rate": 5.652761413450965e-07, "loss": 0.0078, "num_tokens": 1535194556.0, "reward": 0.05078125, "reward_std": 0.08345641195774078, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 1584 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.740234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1894.0, "completions/mean_length": 1662.025390625, "completions/mean_terminated_length": 562.1428833007812, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5410941367244175, "frac_reward_zero_std": 0.8125, "grad_norm": 55.04145002974786, "learning_rate": 5.647403316634181e-07, "loss": 0.0081, "num_tokens": 1536132665.0, "reward": 0.0390625, "reward_std": 0.06657323986291885, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 1585 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2022.0, "completions/mean_length": 1625.16015625, "completions/mean_terminated_length": 544.5694580078125, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.5414355210378083, "frac_reward_zero_std": 0.78125, "grad_norm": 13.457311670723753, "learning_rate": 5.64204501060535e-07, "loss": -0.006, "num_tokens": 1537043931.0, "reward": 0.03515625, "reward_std": 0.07537011057138443, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 1586 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.763671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1856.0, "completions/mean_length": 1701.34765625, "completions/mean_terminated_length": 581.6445922851562, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5417769053511992, "frac_reward_zero_std": 0.75, "grad_norm": 22.038002968898347, "learning_rate": 5.636686502969606e-07, "loss": 0.0125, "num_tokens": 1537998781.0, "reward": 0.068359375, "reward_std": 0.10313399136066437, "rewards/accuracy_reward/mean": 0.068359375, "rewards/accuracy_reward/std": 0.25260838866233826, "step": 1587 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1931.0, "completions/mean_length": 1700.55859375, "completions/mean_terminated_length": 565.5833740234375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5421182896645899, "frac_reward_zero_std": 0.71875, "grad_norm": 16.900784099906076, "learning_rate": 5.631327801332373e-07, "loss": 0.0103, "num_tokens": 1538949067.0, "reward": 0.04296875, "reward_std": 0.09150753915309906, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 1588 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2043.0, "completions/mean_length": 1580.884765625, "completions/mean_terminated_length": 553.2312622070312, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5424596739779807, "frac_reward_zero_std": 0.8125, "grad_norm": 0.8248263849383559, "learning_rate": 5.625968913299344e-07, "loss": 0.0309, "num_tokens": 1539833936.0, "reward": 0.056640625, "reward_std": 0.05519992858171463, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 1589 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.783203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1995.0, "completions/mean_length": 1768.27734375, "completions/mean_terminated_length": 757.7477416992188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5428010582913715, "frac_reward_zero_std": 0.78125, "grad_norm": 9.589589780614224, "learning_rate": 5.620609846476486e-07, "loss": 0.0122, "num_tokens": 1540814574.0, "reward": 0.078125, "reward_std": 0.09951501339673996, "rewards/accuracy_reward/mean": 0.078125, "rewards/accuracy_reward/std": 0.26863065361976624, "step": 1590 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1952.0, "completions/mean_length": 1703.974609375, "completions/mean_terminated_length": 627.508056640625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.5431424426047623, "frac_reward_zero_std": 0.78125, "grad_norm": 11.670179818111938, "learning_rate": 5.615250608470009e-07, "loss": 0.0096, "num_tokens": 1541764561.0, "reward": 0.041015625, "reward_std": 0.07586899399757385, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 1591 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.67578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2001.0, "completions/mean_length": 1579.947265625, "completions/mean_terminated_length": 604.367431640625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5434838269181531, "frac_reward_zero_std": 0.84375, "grad_norm": 56.247460021244855, "learning_rate": 5.609891206886373e-07, "loss": 0.0217, "num_tokens": 1542656710.0, "reward": 0.02734375, "reward_std": 0.05864075943827629, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 1592 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.681640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1923.0, "completions/mean_length": 1609.7421875, "completions/mean_terminated_length": 671.386474609375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5438252112315439, "frac_reward_zero_std": 0.75, "grad_norm": 40.629658863616285, "learning_rate": 5.604531649332267e-07, "loss": 0.0226, "num_tokens": 1543558114.0, "reward": 0.0703125, "reward_std": 0.0969996377825737, "rewards/accuracy_reward/mean": 0.07258064299821854, "rewards/accuracy_reward/std": 0.25970885157585144, "step": 1593 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.759765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1919.0, "completions/mean_length": 1722.44140625, "completions/mean_terminated_length": 693.3251953125, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.5441665955449347, "frac_reward_zero_std": 0.90625, "grad_norm": 2.9854454469198055, "learning_rate": 5.599171943414605e-07, "loss": 0.0049, "num_tokens": 1544520500.0, "reward": 0.029296875, "reward_std": 0.03878315910696983, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 1594 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.837890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1909.0, "completions/mean_length": 1819.775390625, "completions/mean_terminated_length": 640.1566162109375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5445079798583256, "frac_reward_zero_std": 0.8125, "grad_norm": 17.56288186385348, "learning_rate": 5.593812096740507e-07, "loss": 0.0126, "num_tokens": 1545532625.0, "reward": 0.02734375, "reward_std": 0.06116959825158119, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 1595 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.794921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1695.0, "completions/mean_length": 1748.48828125, "completions/mean_terminated_length": 587.5238037109375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.5448493641717163, "frac_reward_zero_std": 0.8125, "grad_norm": 2.1949074880247674, "learning_rate": 5.588452116917299e-07, "loss": 0.0339, "num_tokens": 1546513803.0, "reward": 0.044921875, "reward_std": 0.07581022381782532, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 1596 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.791015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1783.0, "completions/mean_length": 1742.146484375, "completions/mean_terminated_length": 584.4766235351562, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5451907484851071, "frac_reward_zero_std": 0.78125, "grad_norm": 19.177997886752237, "learning_rate": 5.583092011552487e-07, "loss": 0.0193, "num_tokens": 1547483542.0, "reward": 0.05859375, "reward_std": 0.08912044763565063, "rewards/accuracy_reward/mean": 0.05859375, "rewards/accuracy_reward/std": 0.23509246110916138, "step": 1597 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.837890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1959.0, "completions/mean_length": 1815.357421875, "completions/mean_terminated_length": 612.903564453125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5455321327984979, "frac_reward_zero_std": 0.96875, "grad_norm": 10.206085340401652, "learning_rate": 5.577731788253762e-07, "loss": -0.001, "num_tokens": 1548494941.0, "reward": 0.001953125, "reward_std": 0.0078125, "rewards/accuracy_reward/mean": 0.001953125, "rewards/accuracy_reward/std": 0.04419417306780815, "step": 1598 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.826171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2014.0, "completions/mean_length": 1805.357421875, "completions/mean_terminated_length": 652.1235961914062, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5458735171118887, "frac_reward_zero_std": 0.6875, "grad_norm": 59.47086710723055, "learning_rate": 5.572371454628981e-07, "loss": 0.0299, "num_tokens": 1549495204.0, "reward": 0.052734375, "reward_std": 0.09903506934642792, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 1599 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.77734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1608.0, "completions/mean_length": 1730.212890625, "completions/mean_terminated_length": 620.74560546875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5462149014252795, "frac_reward_zero_std": 0.8125, "grad_norm": 24.80823724552991, "learning_rate": 5.567011018286159e-07, "loss": -0.004, "num_tokens": 1550452097.0, "reward": 0.03125, "reward_std": 0.06469620764255524, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 1600 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.78515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2002.0, "completions/mean_length": 1767.408203125, "completions/mean_terminated_length": 741.9727172851562, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.5465562857386703, "frac_reward_zero_std": 0.9375, "grad_norm": 0.29984998692511533, "learning_rate": 5.56165048683345e-07, "loss": 0.0038, "num_tokens": 1551434146.0, "reward": 0.00390625, "reward_std": 0.015625, "rewards/accuracy_reward/mean": 0.004032257944345474, "rewards/accuracy_reward/std": 0.06343589723110199, "step": 1601 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.814453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1761.0, "completions/mean_length": 1779.69921875, "completions/mean_terminated_length": 602.0, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5468976700520611, "frac_reward_zero_std": 0.8125, "grad_norm": 48.14006881714492, "learning_rate": 5.556289867879155e-07, "loss": -0.0051, "num_tokens": 1552420440.0, "reward": 0.0234375, "reward_std": 0.06068411096930504, "rewards/accuracy_reward/mean": 0.024193547666072845, "rewards/accuracy_reward/std": 0.15380479395389557, "step": 1602 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.771484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2018.0, "completions/mean_length": 1731.076171875, "completions/mean_terminated_length": 661.1196899414062, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.547239054365452, "frac_reward_zero_std": 0.75, "grad_norm": 109.44794937858974, "learning_rate": 5.550929169031685e-07, "loss": 0.0179, "num_tokens": 1553385519.0, "reward": 0.0703125, "reward_std": 0.10007822513580322, "rewards/accuracy_reward/mean": 0.0703125, "rewards/accuracy_reward/std": 0.25592297315597534, "step": 1603 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1879.0, "completions/mean_length": 1781.2890625, "completions/mean_terminated_length": 625.5416870117188, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.5475804386788428, "frac_reward_zero_std": 0.8125, "grad_norm": 46.90944229688171, "learning_rate": 5.545568397899575e-07, "loss": -0.0068, "num_tokens": 1554380051.0, "reward": 0.044921875, "reward_std": 0.05452118441462517, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 1604 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.82421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1713.0, "completions/mean_length": 1799.666015625, "completions/mean_terminated_length": 635.2555541992188, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.5479218229922335, "frac_reward_zero_std": 0.78125, "grad_norm": 3.121954078803187, "learning_rate": 5.540207562091459e-07, "loss": 0.0218, "num_tokens": 1555373320.0, "reward": 0.01953125, "reward_std": 0.06233368441462517, "rewards/accuracy_reward/mean": 0.01953125, "rewards/accuracy_reward/std": 0.1385180652141571, "step": 1605 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.787109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1905.0, "completions/mean_length": 1731.94140625, "completions/mean_terminated_length": 563.3944702148438, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5482632073056243, "frac_reward_zero_std": 0.75, "grad_norm": 27.86973018059758, "learning_rate": 5.534846669216062e-07, "loss": 0.0098, "num_tokens": 1556330762.0, "reward": 0.0703125, "reward_std": 0.08698301017284393, "rewards/accuracy_reward/mean": 0.0703125, "rewards/accuracy_reward/std": 0.25592297315597534, "step": 1606 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.783203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1961.0, "completions/mean_length": 1770.404296875, "completions/mean_terminated_length": 767.55859375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5486045916190151, "frac_reward_zero_std": 0.875, "grad_norm": 15.462838239081623, "learning_rate": 5.529485726882193e-07, "loss": 0.0112, "num_tokens": 1557307769.0, "reward": 0.015625, "reward_std": 0.040274329483509064, "rewards/accuracy_reward/mean": 0.015625, "rewards/accuracy_reward/std": 0.12414088100194931, "step": 1607 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.685546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2033.0, "completions/mean_length": 1595.517578125, "completions/mean_terminated_length": 609.0496826171875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.5489459759324059, "frac_reward_zero_std": 0.8125, "grad_norm": 43.23379096921182, "learning_rate": 5.524124742698728e-07, "loss": 0.0377, "num_tokens": 1558197474.0, "reward": 0.072265625, "reward_std": 0.08215849101543427, "rewards/accuracy_reward/mean": 0.072265625, "rewards/accuracy_reward/std": 0.2591804563999176, "step": 1608 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.837890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1863.0, "completions/mean_length": 1834.18359375, "completions/mean_terminated_length": 729.6867065429688, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5492873602457967, "frac_reward_zero_std": 0.84375, "grad_norm": 16.26356002323211, "learning_rate": 5.518763724274602e-07, "loss": 0.0135, "num_tokens": 1559214480.0, "reward": 0.01953125, "reward_std": 0.05001020431518555, "rewards/accuracy_reward/mean": 0.01953125, "rewards/accuracy_reward/std": 0.1385180652141571, "step": 1609 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.77734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2004.0, "completions/mean_length": 1747.224609375, "completions/mean_terminated_length": 697.1491088867188, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.5496287445591875, "frac_reward_zero_std": 0.875, "grad_norm": 5.884596128847715, "learning_rate": 5.513402679218801e-07, "loss": 0.0099, "num_tokens": 1560183795.0, "reward": 0.01171875, "reward_std": 0.03697281330823898, "rewards/accuracy_reward/mean": 0.01171875, "rewards/accuracy_reward/std": 0.10772226005792618, "step": 1610 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.791015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1989.0, "completions/mean_length": 1764.470703125, "completions/mean_terminated_length": 691.299072265625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.5499701288725783, "frac_reward_zero_std": 0.90625, "grad_norm": 75.08828403766782, "learning_rate": 5.50804161514035e-07, "loss": 0.0036, "num_tokens": 1561163604.0, "reward": 0.017578125, "reward_std": 0.03411140665411949, "rewards/accuracy_reward/mean": 0.017578125, "rewards/accuracy_reward/std": 0.13154059648513794, "step": 1611 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1945.0, "completions/mean_length": 1718.046875, "completions/mean_terminated_length": 685.6128540039062, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.5503115131859692, "frac_reward_zero_std": 0.75, "grad_norm": 37.928757065416605, "learning_rate": 5.502680539648296e-07, "loss": 0.008, "num_tokens": 1562116748.0, "reward": 0.056640625, "reward_std": 0.09930095076560974, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 1612 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.76953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1874.0, "completions/mean_length": 1721.806640625, "completions/mean_terminated_length": 632.6525268554688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5506528974993599, "frac_reward_zero_std": 0.8125, "grad_norm": 10.95754696095407, "learning_rate": 5.497319460351706e-07, "loss": 0.0315, "num_tokens": 1563072985.0, "reward": 0.021484375, "reward_std": 0.05782270431518555, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 1613 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.70703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2029.0, "completions/mean_length": 1664.958984375, "completions/mean_terminated_length": 740.5533447265625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5509942818127507, "frac_reward_zero_std": 0.6875, "grad_norm": 60.86552428937514, "learning_rate": 5.491958384859652e-07, "loss": 0.0105, "num_tokens": 1564001844.0, "reward": 0.0625, "reward_std": 0.10628747940063477, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.2422981858253479, "step": 1614 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.748046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1708.05078125, "completions/mean_terminated_length": 698.7442016601562, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.5513356661261415, "frac_reward_zero_std": 0.875, "grad_norm": 8.801982869561556, "learning_rate": 5.486597320781199e-07, "loss": 0.015, "num_tokens": 1564950782.0, "reward": 0.01953125, "reward_std": 0.04412011057138443, "rewards/accuracy_reward/mean": 0.02016128972172737, "rewards/accuracy_reward/std": 0.14069372415542603, "step": 1615 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.685546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1839.0, "completions/mean_length": 1617.49609375, "completions/mean_terminated_length": 678.944091796875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5516770504395323, "frac_reward_zero_std": 0.8125, "grad_norm": 16.429400105017464, "learning_rate": 5.481236275725398e-07, "loss": 0.0102, "num_tokens": 1565853468.0, "reward": 0.056640625, "reward_std": 0.07780590653419495, "rewards/accuracy_reward/mean": 0.058467742055654526, "rewards/accuracy_reward/std": 0.23486268520355225, "step": 1616 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.69921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1771.0, "completions/mean_length": 1607.841796875, "completions/mean_terminated_length": 584.6168823242188, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.5520184347529231, "frac_reward_zero_std": 0.78125, "grad_norm": 14.172403768891147, "learning_rate": 5.475875257301274e-07, "loss": 0.0107, "num_tokens": 1566756187.0, "reward": 0.02734375, "reward_std": 0.07284127175807953, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 1617 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1896.0, "completions/mean_length": 1700.220703125, "completions/mean_terminated_length": 738.7132568359375, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.5523598190663139, "frac_reward_zero_std": 0.84375, "grad_norm": 15.030460260998021, "learning_rate": 5.470514273117807e-07, "loss": 0.0153, "num_tokens": 1567707676.0, "reward": 0.015625, "reward_std": 0.047646719962358475, "rewards/accuracy_reward/mean": 0.015625, "rewards/accuracy_reward/std": 0.12414088100194931, "step": 1618 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.712890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2044.0, "completions/mean_length": 1654.228515625, "completions/mean_terminated_length": 676.49658203125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5527012033797047, "frac_reward_zero_std": 0.78125, "grad_norm": 26.41145164261852, "learning_rate": 5.465153330783939e-07, "loss": 0.0205, "num_tokens": 1568627313.0, "reward": 0.068359375, "reward_std": 0.08864613622426987, "rewards/accuracy_reward/mean": 0.068359375, "rewards/accuracy_reward/std": 0.25260838866233826, "step": 1619 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2041.0, "completions/mean_length": 1748.0703125, "completions/mean_terminated_length": 626.1111450195312, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5530425876930956, "frac_reward_zero_std": 0.78125, "grad_norm": 15.904342611748271, "learning_rate": 5.459792437908542e-07, "loss": 0.0003, "num_tokens": 1569596293.0, "reward": 0.056640625, "reward_std": 0.0891871303319931, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 1620 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.74609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1816.0, "completions/mean_length": 1706.689453125, "completions/mean_terminated_length": 703.7615356445312, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5533839720064863, "frac_reward_zero_std": 0.71875, "grad_norm": 9.920335216339595, "learning_rate": 5.454431602100425e-07, "loss": 0.0223, "num_tokens": 1570543638.0, "reward": 0.0390625, "reward_std": 0.09005707502365112, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 1621 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.72265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1970.0, "completions/mean_length": 1643.3203125, "completions/mean_terminated_length": 588.8732299804688, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.5537253563198771, "frac_reward_zero_std": 0.8125, "grad_norm": 21.087527788184534, "learning_rate": 5.449070830968316e-07, "loss": 0.0275, "num_tokens": 1571461706.0, "reward": 0.041015625, "reward_std": 0.069866843521595, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 1622 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.751953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1726.0, "completions/mean_length": 1709.193359375, "completions/mean_terminated_length": 682.1023559570312, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.5540667406332679, "frac_reward_zero_std": 0.84375, "grad_norm": 18.13806863823379, "learning_rate": 5.443710132120846e-07, "loss": 0.0134, "num_tokens": 1572413085.0, "reward": 0.03125, "reward_std": 0.05754890665411949, "rewards/accuracy_reward/mean": 0.032258063554763794, "rewards/accuracy_reward/std": 0.17686307430267334, "step": 1623 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1857.0, "completions/mean_length": 1656.2734375, "completions/mean_terminated_length": 692.8378295898438, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5544081249466587, "frac_reward_zero_std": 0.875, "grad_norm": 28.706270164835576, "learning_rate": 5.438349513166549e-07, "loss": -0.0039, "num_tokens": 1573333689.0, "reward": 0.021484375, "reward_std": 0.04230976849794388, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 1624 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.693359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1954.0, "completions/mean_length": 1643.267578125, "completions/mean_terminated_length": 728.1082763671875, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.5547495092600495, "frac_reward_zero_std": 0.6875, "grad_norm": 33.42744635694024, "learning_rate": 5.432988981713842e-07, "loss": 0.0383, "num_tokens": 1574246178.0, "reward": 0.087890625, "reward_std": 0.1285536289215088, "rewards/accuracy_reward/mean": 0.087890625, "rewards/accuracy_reward/std": 0.2834126651287079, "step": 1625 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1712.0, "completions/mean_length": 1713.986328125, "completions/mean_terminated_length": 622.8750610351562, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.5550908935734403, "frac_reward_zero_std": 0.65625, "grad_norm": 37.37660652247251, "learning_rate": 5.42762854537102e-07, "loss": 0.0296, "num_tokens": 1575198587.0, "reward": 0.048828125, "reward_std": 0.11779527366161346, "rewards/accuracy_reward/mean": 0.05040322616696358, "rewards/accuracy_reward/std": 0.21899642050266266, "step": 1626 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.677734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2044.0, "completions/mean_length": 1621.921875, "completions/mean_terminated_length": 726.13330078125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5554322778868311, "frac_reward_zero_std": 0.84375, "grad_norm": 61.239165904831594, "learning_rate": 5.422268211746239e-07, "loss": 0.0057, "num_tokens": 1576111059.0, "reward": 0.056640625, "reward_std": 0.07447256147861481, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 1627 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.685546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2029.0, "completions/mean_length": 1593.0859375, "completions/mean_terminated_length": 601.3167724609375, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.555773662200222, "frac_reward_zero_std": 0.71875, "grad_norm": 18.74271606092731, "learning_rate": 5.416907988447514e-07, "loss": 0.0039, "num_tokens": 1577012991.0, "reward": 0.0703125, "reward_std": 0.1190669909119606, "rewards/accuracy_reward/mean": 0.0703125, "rewards/accuracy_reward/std": 0.25592297315597534, "step": 1628 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.70703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1977.0, "completions/mean_length": 1613.962890625, "completions/mean_terminated_length": 566.4866943359375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.5561150465136127, "frac_reward_zero_std": 0.71875, "grad_norm": 36.865221079727164, "learning_rate": 5.411547883082701e-07, "loss": 0.0189, "num_tokens": 1577913356.0, "reward": 0.056640625, "reward_std": 0.10854585468769073, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 1629 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1834.0, "completions/mean_length": 1653.994140625, "completions/mean_terminated_length": 647.0902709960938, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5564564308270035, "frac_reward_zero_std": 0.8125, "grad_norm": 7.796400320146339, "learning_rate": 5.406187903259491e-07, "loss": 0.01, "num_tokens": 1578831897.0, "reward": 0.046875, "reward_std": 0.07053204625844955, "rewards/accuracy_reward/mean": 0.04838709533214569, "rewards/accuracy_reward/std": 0.21479946374893188, "step": 1630 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.708984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1964.0, "completions/mean_length": 1661.341796875, "completions/mean_terminated_length": 719.3489990234375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.5567978151403943, "frac_reward_zero_std": 0.84375, "grad_norm": 7.989295862922379, "learning_rate": 5.400828056585394e-07, "loss": 0.0047, "num_tokens": 1579770696.0, "reward": 0.013671875, "reward_std": 0.04478531330823898, "rewards/accuracy_reward/mean": 0.013671875, "rewards/accuracy_reward/std": 0.1162383034825325, "step": 1631 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.728515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1953.0, "completions/mean_length": 1693.296875, "completions/mean_terminated_length": 741.4676513671875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5571391994537851, "frac_reward_zero_std": 0.84375, "grad_norm": 75.67640237032624, "learning_rate": 5.395468350667732e-07, "loss": -0.0002, "num_tokens": 1580717920.0, "reward": 0.025390625, "reward_std": 0.05335709825158119, "rewards/accuracy_reward/mean": 0.025390625, "rewards/accuracy_reward/std": 0.15746226906776428, "step": 1632 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1867.0, "completions/mean_length": 1624.634765625, "completions/mean_terminated_length": 658.4935913085938, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5574805837671759, "frac_reward_zero_std": 0.78125, "grad_norm": 40.33398414718872, "learning_rate": 5.390108793113628e-07, "loss": -0.0025, "num_tokens": 1581620837.0, "reward": 0.044921875, "reward_std": 0.08537977933883667, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 1633 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.720703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1912.0, "completions/mean_length": 1661.6796875, "completions/mean_terminated_length": 664.8111572265625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5578219680805667, "frac_reward_zero_std": 0.65625, "grad_norm": 23.260771731586303, "learning_rate": 5.384749391529991e-07, "loss": 0.0143, "num_tokens": 1582551809.0, "reward": 0.103515625, "reward_std": 0.1405133157968521, "rewards/accuracy_reward/mean": 0.103515625, "rewards/accuracy_reward/std": 0.30492907762527466, "step": 1634 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1941.0, "completions/mean_length": 1664.76171875, "completions/mean_terminated_length": 685.375, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.5581633523939575, "frac_reward_zero_std": 0.75, "grad_norm": 10.739587039328736, "learning_rate": 5.379390153523515e-07, "loss": 0.0126, "num_tokens": 1583481159.0, "reward": 0.041015625, "reward_std": 0.0879673957824707, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 1635 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.75390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2003.0, "completions/mean_length": 1729.625, "completions/mean_terminated_length": 754.698486328125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.5585047367073483, "frac_reward_zero_std": 0.78125, "grad_norm": 30.268747247832582, "learning_rate": 5.374031086700654e-07, "loss": 0.0046, "num_tokens": 1584448023.0, "reward": 0.037109375, "reward_std": 0.07575597614049911, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 1636 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.66796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1967.0, "completions/mean_length": 1612.603515625, "completions/mean_terminated_length": 736.941162109375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.558846121020739, "frac_reward_zero_std": 0.875, "grad_norm": 2.6913769660511266, "learning_rate": 5.368672198667627e-07, "loss": 0.0044, "num_tokens": 1585360428.0, "reward": 0.025390625, "reward_std": 0.04230976849794388, "rewards/accuracy_reward/mean": 0.025390625, "rewards/accuracy_reward/std": 0.15746226906776428, "step": 1637 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1939.0, "completions/mean_length": 1655.478515625, "completions/mean_terminated_length": 759.724365234375, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.5591875053341299, "frac_reward_zero_std": 0.875, "grad_norm": 1.6864334362397269, "learning_rate": 5.363313497030395e-07, "loss": 0.0054, "num_tokens": 1586289313.0, "reward": 0.03515625, "reward_std": 0.04505911096930504, "rewards/accuracy_reward/mean": 0.03629032149910927, "rewards/accuracy_reward/std": 0.1872003972530365, "step": 1638 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.650390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1947.0, "completions/mean_length": 1543.3359375, "completions/mean_terminated_length": 604.4915771484375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.5595288896475207, "frac_reward_zero_std": 0.78125, "grad_norm": 28.75318906139253, "learning_rate": 5.357954989394651e-07, "loss": -0.0048, "num_tokens": 1587148029.0, "reward": 0.064453125, "reward_std": 0.08648413419723511, "rewards/accuracy_reward/mean": 0.06653226166963577, "rewards/accuracy_reward/std": 0.24946178495883942, "step": 1639 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.654296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2043.0, "completions/mean_length": 1583.01953125, "completions/mean_terminated_length": 702.9717407226562, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5598702739609115, "frac_reward_zero_std": 0.78125, "grad_norm": 5.467112168998574, "learning_rate": 5.35259668336582e-07, "loss": 0.009, "num_tokens": 1588032471.0, "reward": 0.037109375, "reward_std": 0.07763300836086273, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 1640 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1958.0, "completions/mean_length": 1599.435546875, "completions/mean_terminated_length": 681.2738037109375, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.5602116582743023, "frac_reward_zero_std": 0.6875, "grad_norm": 18.105598438490166, "learning_rate": 5.347238586549036e-07, "loss": 0.0256, "num_tokens": 1588930006.0, "reward": 0.083984375, "reward_std": 0.12064876407384872, "rewards/accuracy_reward/mean": 0.083984375, "rewards/accuracy_reward/std": 0.2776356339454651, "step": 1641 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.75, "completions/max_length": 2048.0, "completions/max_terminated_length": 2004.0, "completions/mean_length": 1687.611328125, "completions/mean_terminated_length": 606.4453125, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.5605530425876931, "frac_reward_zero_std": 0.78125, "grad_norm": 10.322341414725255, "learning_rate": 5.341880706549138e-07, "loss": 0.043, "num_tokens": 1589871519.0, "reward": 0.03515625, "reward_std": 0.06822281330823898, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 1642 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1877.0, "completions/mean_length": 1605.12890625, "completions/mean_terminated_length": 556.2236938476562, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5608944269010839, "frac_reward_zero_std": 0.71875, "grad_norm": 58.135449586735845, "learning_rate": 5.336523050970657e-07, "loss": 0.0217, "num_tokens": 1590773553.0, "reward": 0.052734375, "reward_std": 0.10057821124792099, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 1643 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.65625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2021.0, "completions/mean_length": 1556.451171875, "completions/mean_terminated_length": 618.039794921875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5612358112144747, "frac_reward_zero_std": 0.8125, "grad_norm": 4.517399172075493, "learning_rate": 5.331165627417807e-07, "loss": 0.0098, "num_tokens": 1591653128.0, "reward": 0.021484375, "reward_std": 0.0546875, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 1644 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.689453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2031.0, "completions/mean_length": 1622.578125, "completions/mean_terminated_length": 678.0880126953125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5615771955278654, "frac_reward_zero_std": 0.6875, "grad_norm": 29.488694737350862, "learning_rate": 5.325808443494467e-07, "loss": 0.007, "num_tokens": 1592568304.0, "reward": 0.1015625, "reward_std": 0.12225200235843658, "rewards/accuracy_reward/mean": 0.1015625, "rewards/accuracy_reward/std": 0.30236753821372986, "step": 1645 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.673828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1937.0, "completions/mean_length": 1551.8671875, "completions/mean_terminated_length": 526.9221801757812, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5619185798412563, "frac_reward_zero_std": 0.78125, "grad_norm": 19.09939734373746, "learning_rate": 5.32045150680418e-07, "loss": 0.0195, "num_tokens": 1593433500.0, "reward": 0.037109375, "reward_std": 0.07003316283226013, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 1646 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.583984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1985.0, "completions/mean_length": 1536.521484375, "completions/mean_terminated_length": 818.530517578125, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.5622599641546471, "frac_reward_zero_std": 0.625, "grad_norm": 23.450097969172788, "learning_rate": 5.31509482495014e-07, "loss": 0.0199, "num_tokens": 1594292695.0, "reward": 0.08203125, "reward_std": 0.1348828822374344, "rewards/accuracy_reward/mean": 0.08203125, "rewards/accuracy_reward/std": 0.2746807038784027, "step": 1647 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.630859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1539.884765625, "completions/mean_terminated_length": 671.74072265625, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.5626013484680379, "frac_reward_zero_std": 0.65625, "grad_norm": 18.61825255663601, "learning_rate": 5.309738405535177e-07, "loss": 0.0293, "num_tokens": 1595159292.0, "reward": 0.10546875, "reward_std": 0.12310365587472916, "rewards/accuracy_reward/mean": 0.10546875, "rewards/accuracy_reward/std": 0.3074568510055542, "step": 1648 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.705078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1826.0, "completions/mean_length": 1637.900390625, "completions/mean_terminated_length": 657.4635620117188, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5629427327814287, "frac_reward_zero_std": 0.75, "grad_norm": 24.422153094050394, "learning_rate": 5.304382256161746e-07, "loss": 0.0262, "num_tokens": 1596071241.0, "reward": 0.0703125, "reward_std": 0.10351984947919846, "rewards/accuracy_reward/mean": 0.0703125, "rewards/accuracy_reward/std": 0.25592297315597534, "step": 1649 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.708984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1963.0, "completions/mean_length": 1639.8203125, "completions/mean_terminated_length": 645.39599609375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5632841170948195, "frac_reward_zero_std": 0.8125, "grad_norm": 15.547314072867204, "learning_rate": 5.299026384431929e-07, "loss": 0.0296, "num_tokens": 1596984077.0, "reward": 0.07421875, "reward_std": 0.08482664078474045, "rewards/accuracy_reward/mean": 0.07421875, "rewards/accuracy_reward/std": 0.2623828947544098, "step": 1650 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.5859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1979.0, "completions/mean_length": 1463.548828125, "completions/mean_terminated_length": 636.4953002929688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5636255014082103, "frac_reward_zero_std": 0.6875, "grad_norm": 32.958100487220385, "learning_rate": 5.293670797947396e-07, "loss": 0.0314, "num_tokens": 1597803286.0, "reward": 0.076171875, "reward_std": 0.12246465682983398, "rewards/accuracy_reward/mean": 0.076171875, "rewards/accuracy_reward/std": 0.26553234457969666, "step": 1651 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.740234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2011.0, "completions/mean_length": 1692.24609375, "completions/mean_terminated_length": 678.481201171875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.5639668857216011, "frac_reward_zero_std": 0.75, "grad_norm": 46.8267009918828, "learning_rate": 5.288315504309429e-07, "loss": -0.0072, "num_tokens": 1598750116.0, "reward": 0.029296875, "reward_std": 0.07768725603818893, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 1652 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.701171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.0, "completions/mean_length": 1647.994140625, "completions/mean_terminated_length": 709.4183349609375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.5643082700349918, "frac_reward_zero_std": 0.78125, "grad_norm": 36.00313758741006, "learning_rate": 5.282960511118882e-07, "loss": 0.0127, "num_tokens": 1599670721.0, "reward": 0.041015625, "reward_std": 0.0814797431230545, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 1653 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.64453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1911.0, "completions/mean_length": 1527.19921875, "completions/mean_terminated_length": 582.89013671875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.5646496543483827, "frac_reward_zero_std": 0.84375, "grad_norm": 51.006675241895266, "learning_rate": 5.277605825976195e-07, "loss": 0.0115, "num_tokens": 1600531223.0, "reward": 0.03515625, "reward_std": 0.06672705709934235, "rewards/accuracy_reward/mean": 0.03629032149910927, "rewards/accuracy_reward/std": 0.1872003823518753, "step": 1654 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.75390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1898.0, "completions/mean_length": 1736.5, "completions/mean_terminated_length": 782.2222900390625, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.5649910386617735, "frac_reward_zero_std": 0.78125, "grad_norm": 110.9152983650968, "learning_rate": 5.272251456481363e-07, "loss": -0.0005, "num_tokens": 1601493335.0, "reward": 0.037109375, "reward_std": 0.07575597614049911, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 1655 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.5546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1900.0, "completions/mean_length": 1440.62109375, "completions/mean_terminated_length": 684.0614013671875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5653324229751643, "frac_reward_zero_std": 0.78125, "grad_norm": 4.6806848587855985, "learning_rate": 5.266897410233934e-07, "loss": 0.0266, "num_tokens": 1602306357.0, "reward": 0.029296875, "reward_std": 0.0723423957824707, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 1656 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1852.0, "completions/mean_length": 1685.63671875, "completions/mean_terminated_length": 684.1102905273438, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.5656738072885551, "frac_reward_zero_std": 0.75, "grad_norm": 79.23697229736925, "learning_rate": 5.261543694832994e-07, "loss": 0.0149, "num_tokens": 1603242123.0, "reward": 0.04296875, "reward_std": 0.08655644953250885, "rewards/accuracy_reward/mean": 0.04435483738780022, "rewards/accuracy_reward/std": 0.2060900777578354, "step": 1657 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.744140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1984.0, "completions/mean_length": 1705.181640625, "completions/mean_terminated_length": 708.1297607421875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.5660151916019459, "frac_reward_zero_std": 0.75, "grad_norm": 22.73665048768005, "learning_rate": 5.256190317877164e-07, "loss": 0.0114, "num_tokens": 1604197016.0, "reward": 0.044921875, "reward_std": 0.0862836092710495, "rewards/accuracy_reward/mean": 0.0463709682226181, "rewards/accuracy_reward/std": 0.21049949526786804, "step": 1658 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.705078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1893.0, "completions/mean_length": 1647.029296875, "completions/mean_terminated_length": 688.417236328125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.5663565759153367, "frac_reward_zero_std": 0.8125, "grad_norm": 10.532580196132601, "learning_rate": 5.250837286964585e-07, "loss": 0.0146, "num_tokens": 1605120727.0, "reward": 0.0234375, "reward_std": 0.062167368829250336, "rewards/accuracy_reward/mean": 0.0234375, "rewards/accuracy_reward/std": 0.15143637359142303, "step": 1659 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.619140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1860.0, "completions/mean_length": 1530.748046875, "completions/mean_terminated_length": 689.882080078125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.5666979602287275, "frac_reward_zero_std": 0.90625, "grad_norm": 6.574757140261458, "learning_rate": 5.245484609692906e-07, "loss": 0.0037, "num_tokens": 1605977958.0, "reward": 0.015625, "reward_std": 0.03125, "rewards/accuracy_reward/mean": 0.015625, "rewards/accuracy_reward/std": 0.12414088100194931, "step": 1660 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.69921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2043.0, "completions/mean_length": 1622.96875, "completions/mean_terminated_length": 634.9091186523438, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.5670393445421182, "frac_reward_zero_std": 0.65625, "grad_norm": 42.688717085118896, "learning_rate": 5.240132293659268e-07, "loss": 0.0357, "num_tokens": 1606890838.0, "reward": 0.10546875, "reward_std": 0.13172858953475952, "rewards/accuracy_reward/mean": 0.10546875, "rewards/accuracy_reward/std": 0.3074568510055542, "step": 1661 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1916.0, "completions/mean_length": 1550.03515625, "completions/mean_terminated_length": 662.3587036132812, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.567380728855509, "frac_reward_zero_std": 0.8125, "grad_norm": 8.403924701857363, "learning_rate": 5.234780346460314e-07, "loss": 0.0094, "num_tokens": 1607759256.0, "reward": 0.052734375, "reward_std": 0.069866843521595, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 1662 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.705078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1834.0, "completions/mean_length": 1643.822265625, "completions/mean_terminated_length": 677.5430297851562, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5677221131688999, "frac_reward_zero_std": 0.8125, "grad_norm": 53.38768495225109, "learning_rate": 5.229428775692146e-07, "loss": 0.025, "num_tokens": 1608682717.0, "reward": 0.052734375, "reward_std": 0.07003316283226013, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 1663 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.689453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1861.0, "completions/mean_length": 1600.130859375, "completions/mean_terminated_length": 606.0880126953125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5680634974822907, "frac_reward_zero_std": 0.6875, "grad_norm": 45.39491460847951, "learning_rate": 5.224077588950342e-07, "loss": 0.0422, "num_tokens": 1609588800.0, "reward": 0.078125, "reward_std": 0.12353022396564484, "rewards/accuracy_reward/mean": 0.078125, "rewards/accuracy_reward/std": 0.26863065361976624, "step": 1664 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.607421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2030.0, "completions/mean_length": 1524.009765625, "completions/mean_terminated_length": 713.2586669921875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5684048817956815, "frac_reward_zero_std": 0.65625, "grad_norm": 195.70172751033917, "learning_rate": 5.218726793829936e-07, "loss": 0.0108, "num_tokens": 1610435173.0, "reward": 0.10546875, "reward_std": 0.13682594895362854, "rewards/accuracy_reward/mean": 0.10546875, "rewards/accuracy_reward/std": 0.3074568510055542, "step": 1665 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.599609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1998.0, "completions/mean_length": 1507.548828125, "completions/mean_terminated_length": 698.1902465820312, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5687462661090723, "frac_reward_zero_std": 0.78125, "grad_norm": 0.935269412343542, "learning_rate": 5.213376397925399e-07, "loss": 0.0265, "num_tokens": 1611281678.0, "reward": 0.03125, "reward_std": 0.07708083093166351, "rewards/accuracy_reward/mean": 0.032258063554763794, "rewards/accuracy_reward/std": 0.17686307430267334, "step": 1666 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.642578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2015.0, "completions/mean_length": 1558.541015625, "completions/mean_terminated_length": 678.5846557617188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5690876504224631, "frac_reward_zero_std": 0.71875, "grad_norm": 48.18749015750313, "learning_rate": 5.208026408830641e-07, "loss": 0.0147, "num_tokens": 1612154035.0, "reward": 0.05859375, "reward_std": 0.10915680229663849, "rewards/accuracy_reward/mean": 0.05859375, "rewards/accuracy_reward/std": 0.23509246110916138, "step": 1667 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.65625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1838.0, "completions/mean_length": 1548.416015625, "completions/mean_terminated_length": 594.8806762695312, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.5694290347358539, "frac_reward_zero_std": 0.84375, "grad_norm": 33.075900760294196, "learning_rate": 5.202676834138993e-07, "loss": 0.0007, "num_tokens": 1613033112.0, "reward": 0.03125, "reward_std": 0.05695698410272598, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 1668 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.630859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1868.0, "completions/mean_length": 1565.953125, "completions/mean_terminated_length": 742.1375732421875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5697704190492446, "frac_reward_zero_std": 0.71875, "grad_norm": 43.26088067331689, "learning_rate": 5.1973276814432e-07, "loss": 0.0195, "num_tokens": 1613906848.0, "reward": 0.091796875, "reward_std": 0.12340601533651352, "rewards/accuracy_reward/mean": 0.091796875, "rewards/accuracy_reward/std": 0.289021372795105, "step": 1669 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.630859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2037.0, "completions/mean_length": 1548.76953125, "completions/mean_terminated_length": 695.5872802734375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5701118033626354, "frac_reward_zero_std": 0.71875, "grad_norm": 5.572669498956679, "learning_rate": 5.191978958335402e-07, "loss": 0.0227, "num_tokens": 1614779066.0, "reward": 0.08203125, "reward_std": 0.11268904060125351, "rewards/accuracy_reward/mean": 0.08203125, "rewards/accuracy_reward/std": 0.2746807038784027, "step": 1670 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.673828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1835.0, "completions/mean_length": 1599.775390625, "completions/mean_terminated_length": 673.8024291992188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5704531876760263, "frac_reward_zero_std": 0.65625, "grad_norm": 19.083995537021174, "learning_rate": 5.186630672407133e-07, "loss": 0.0186, "num_tokens": 1615675767.0, "reward": 0.064453125, "reward_std": 0.12303133308887482, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 1671 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.62890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1872.0, "completions/mean_length": 1549.361328125, "completions/mean_terminated_length": 704.300048828125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5707945719894171, "frac_reward_zero_std": 0.8125, "grad_norm": 12.362695622488507, "learning_rate": 5.181282831249311e-07, "loss": 0.0005, "num_tokens": 1616549216.0, "reward": 0.029296875, "reward_std": 0.06519509106874466, "rewards/accuracy_reward/mean": 0.030241934582591057, "rewards/accuracy_reward/std": 0.1714252382516861, "step": 1672 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.705078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1774.0, "completions/mean_length": 1638.623046875, "completions/mean_terminated_length": 659.9138793945312, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5711359563028079, "frac_reward_zero_std": 0.6875, "grad_norm": 27.431925772868905, "learning_rate": 5.175935442452213e-07, "loss": -0.0097, "num_tokens": 1617471103.0, "reward": 0.056640625, "reward_std": 0.11036626994609833, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 1673 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1973.0, "completions/mean_length": 1580.474609375, "completions/mean_terminated_length": 747.059814453125, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.5714773406161987, "frac_reward_zero_std": 0.71875, "grad_norm": 16.107035256130246, "learning_rate": 5.170588513605485e-07, "loss": 0.027, "num_tokens": 1618355714.0, "reward": 0.04296875, "reward_std": 0.09930886328220367, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 1674 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.65234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1983.0, "completions/mean_length": 1576.763671875, "completions/mean_terminated_length": 692.53369140625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5718187249295895, "frac_reward_zero_std": 0.65625, "grad_norm": 21.902487619837963, "learning_rate": 5.165242052298112e-07, "loss": 0.046, "num_tokens": 1619242505.0, "reward": 0.091796875, "reward_std": 0.12451037019491196, "rewards/accuracy_reward/mean": 0.09791667014360428, "rewards/accuracy_reward/std": 0.2975119948387146, "step": 1675 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2013.0, "completions/mean_length": 1532.61328125, "completions/mean_terminated_length": 701.9387817382812, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.5721601092429803, "frac_reward_zero_std": 0.8125, "grad_norm": 29.906182728225684, "learning_rate": 5.159896066118417e-07, "loss": -0.0075, "num_tokens": 1620108595.0, "reward": 0.044921875, "reward_std": 0.064970001578331, "rewards/accuracy_reward/mean": 0.0463709682226181, "rewards/accuracy_reward/std": 0.21049949526786804, "step": 1676 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.70703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1696.0, "completions/mean_length": 1648.66015625, "completions/mean_terminated_length": 684.9200439453125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.572501493556371, "frac_reward_zero_std": 0.8125, "grad_norm": 2.319766164938501, "learning_rate": 5.15455056265405e-07, "loss": 0.0226, "num_tokens": 1621027445.0, "reward": 0.05078125, "reward_std": 0.07289456576108932, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 1677 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.638671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1977.0, "completions/mean_length": 1557.611328125, "completions/mean_terminated_length": 690.8162231445312, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5728428778697618, "frac_reward_zero_std": 0.75, "grad_norm": 7.976963634444491, "learning_rate": 5.149205549491975e-07, "loss": 0.0261, "num_tokens": 1621899198.0, "reward": 0.03515625, "reward_std": 0.08219823986291885, "rewards/accuracy_reward/mean": 0.03629032149910927, "rewards/accuracy_reward/std": 0.1872003823518753, "step": 1678 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.732421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2030.0, "completions/mean_length": 1702.068359375, "completions/mean_terminated_length": 755.934326171875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.5731842621831527, "frac_reward_zero_std": 0.90625, "grad_norm": 6.006209757059818, "learning_rate": 5.143861034218462e-07, "loss": 0.0027, "num_tokens": 1622855745.0, "reward": 0.01953125, "reward_std": 0.036034777760505676, "rewards/accuracy_reward/mean": 0.01953125, "rewards/accuracy_reward/std": 0.1385180652141571, "step": 1679 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1996.0, "completions/mean_length": 1531.52734375, "completions/mean_terminated_length": 670.7396240234375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.5735256464965435, "frac_reward_zero_std": 0.71875, "grad_norm": 20.07107750501197, "learning_rate": 5.138517024419071e-07, "loss": 0.0318, "num_tokens": 1623714319.0, "reward": 0.048828125, "reward_std": 0.10194281488656998, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 1680 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1992.0, "completions/mean_length": 1528.04296875, "completions/mean_terminated_length": 601.1630249023438, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.5738670308099343, "frac_reward_zero_std": 0.8125, "grad_norm": 14.94345538810307, "learning_rate": 5.13317352767865e-07, "loss": 0.0168, "num_tokens": 1624580997.0, "reward": 0.044921875, "reward_std": 0.07302113622426987, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 1681 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.685546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2013.0, "completions/mean_length": 1615.00390625, "completions/mean_terminated_length": 671.0186157226562, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.5742084151233251, "frac_reward_zero_std": 0.71875, "grad_norm": 12.158901872846986, "learning_rate": 5.127830551581311e-07, "loss": 0.031, "num_tokens": 1625482647.0, "reward": 0.0625, "reward_std": 0.09671792387962341, "rewards/accuracy_reward/mean": 0.06451612710952759, "rewards/accuracy_reward/std": 0.2459181249141693, "step": 1682 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.732421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2024.0, "completions/mean_length": 1698.96875, "completions/mean_terminated_length": 743.9780883789062, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5745497994367159, "frac_reward_zero_std": 0.75, "grad_norm": 37.86543485284138, "learning_rate": 5.122488103710435e-07, "loss": 0.0128, "num_tokens": 1626428183.0, "reward": 0.03125, "reward_std": 0.07069835811853409, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 1683 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.783203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1989.0, "completions/mean_length": 1747.787109375, "completions/mean_terminated_length": 663.2342529296875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5748911837501067, "frac_reward_zero_std": 0.65625, "grad_norm": 40.71383223966807, "learning_rate": 5.117146191648647e-07, "loss": 0.0347, "num_tokens": 1627394986.0, "reward": 0.087890625, "reward_std": 0.1312296986579895, "rewards/accuracy_reward/mean": 0.087890625, "rewards/accuracy_reward/std": 0.2834126651287079, "step": 1684 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.677734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2013.0, "completions/mean_length": 1567.5625, "completions/mean_terminated_length": 557.1878662109375, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.5752325680634974, "frac_reward_zero_std": 0.8125, "grad_norm": 20.638246866731627, "learning_rate": 5.111804822977814e-07, "loss": 0.0017, "num_tokens": 1628271290.0, "reward": 0.0703125, "reward_std": 0.08044223487377167, "rewards/accuracy_reward/mean": 0.0703125, "rewards/accuracy_reward/std": 0.25592297315597534, "step": 1685 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.67578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1848.0, "completions/mean_length": 1568.5859375, "completions/mean_terminated_length": 569.3252563476562, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5755739523768882, "frac_reward_zero_std": 0.78125, "grad_norm": 39.55417597841334, "learning_rate": 5.106464005279034e-07, "loss": 0.0184, "num_tokens": 1629150246.0, "reward": 0.068359375, "reward_std": 0.09126891195774078, "rewards/accuracy_reward/mean": 0.068359375, "rewards/accuracy_reward/std": 0.25260838866233826, "step": 1686 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.720703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2031.0, "completions/mean_length": 1648.943359375, "completions/mean_terminated_length": 619.2097778320312, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5759153366902791, "frac_reward_zero_std": 0.84375, "grad_norm": 10.018973187369816, "learning_rate": 5.101123746132622e-07, "loss": -0.0046, "num_tokens": 1630073945.0, "reward": 0.03125, "reward_std": 0.05641176179051399, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 1687 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.775390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1950.0, "completions/mean_length": 1745.142578125, "completions/mean_terminated_length": 700.5477905273438, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5762567210036699, "frac_reward_zero_std": 0.8125, "grad_norm": 86.75565488956244, "learning_rate": 5.095784053118094e-07, "loss": 0.0092, "num_tokens": 1631060002.0, "reward": 0.02734375, "reward_std": 0.06233368441462517, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 1688 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.732421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2001.0, "completions/mean_length": 1671.2421875, "completions/mean_terminated_length": 639.9708251953125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5765981053170607, "frac_reward_zero_std": 0.6875, "grad_norm": 15.014602420932516, "learning_rate": 5.090444933814171e-07, "loss": 0.026, "num_tokens": 1631997006.0, "reward": 0.0546875, "reward_std": 0.11135159432888031, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 1689 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.75390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1848.0, "completions/mean_length": 1688.90234375, "completions/mean_terminated_length": 588.8095703125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5769394896304515, "frac_reward_zero_std": 0.875, "grad_norm": 24.722525212805873, "learning_rate": 5.085106395798756e-07, "loss": -0.0004, "num_tokens": 1632934188.0, "reward": 0.044921875, "reward_std": 0.052045635879039764, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 1690 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.74609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2044.0, "completions/mean_length": 1694.869140625, "completions/mean_terminated_length": 657.2077026367188, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5772808739438423, "frac_reward_zero_std": 0.84375, "grad_norm": 43.27838630705788, "learning_rate": 5.079768446648926e-07, "loss": 0.0032, "num_tokens": 1633877993.0, "reward": 0.052734375, "reward_std": 0.0742793083190918, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 1691 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.767578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1890.0, "completions/mean_length": 1743.4296875, "completions/mean_terminated_length": 737.5798950195312, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5776222582572331, "frac_reward_zero_std": 0.6875, "grad_norm": 15.678326575986901, "learning_rate": 5.074431093940927e-07, "loss": 0.0424, "num_tokens": 1634841285.0, "reward": 0.05078125, "reward_std": 0.102440744638443, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 1692 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1876.0, "completions/mean_length": 1717.419921875, "completions/mean_terminated_length": 588.887939453125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5779636425706238, "frac_reward_zero_std": 0.78125, "grad_norm": 14.946153532651053, "learning_rate": 5.069094345250152e-07, "loss": 0.0217, "num_tokens": 1635793964.0, "reward": 0.072265625, "reward_std": 0.0868302434682846, "rewards/accuracy_reward/mean": 0.072265625, "rewards/accuracy_reward/std": 0.2591804563999176, "step": 1693 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.76171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1992.0, "completions/mean_length": 1741.654296875, "completions/mean_terminated_length": 762.3524169921875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.5783050268840146, "frac_reward_zero_std": 0.78125, "grad_norm": 21.58139157783209, "learning_rate": 5.063758208151139e-07, "loss": 0.0269, "num_tokens": 1636770795.0, "reward": 0.048828125, "reward_std": 0.08082009106874466, "rewards/accuracy_reward/mean": 0.05040322616696358, "rewards/accuracy_reward/std": 0.21899642050266266, "step": 1694 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.751953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2036.0, "completions/mean_length": 1736.56640625, "completions/mean_terminated_length": 792.4566650390625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5786464111974055, "frac_reward_zero_std": 0.84375, "grad_norm": 18.95436493028647, "learning_rate": 5.058422690217559e-07, "loss": 0.0135, "num_tokens": 1637738013.0, "reward": 0.015625, "reward_std": 0.04670868441462517, "rewards/accuracy_reward/mean": 0.015625, "rewards/accuracy_reward/std": 0.12414088100194931, "step": 1695 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.767578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1966.0, "completions/mean_length": 1716.841796875, "completions/mean_terminated_length": 623.1848754882812, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5789877955107963, "frac_reward_zero_std": 0.8125, "grad_norm": 68.46485994060792, "learning_rate": 5.053087799022207e-07, "loss": 0.014, "num_tokens": 1638687692.0, "reward": 0.029296875, "reward_std": 0.0659080371260643, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 1696 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1889.0, "completions/mean_length": 1778.5, "completions/mean_terminated_length": 610.6666870117188, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.5793291798241871, "frac_reward_zero_std": 0.84375, "grad_norm": 39.97449024934502, "learning_rate": 5.047753542136981e-07, "loss": 0.0034, "num_tokens": 1639682988.0, "reward": 0.021484375, "reward_std": 0.04973640665411949, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 1697 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.791015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1831.0, "completions/mean_length": 1752.87890625, "completions/mean_terminated_length": 636.3644409179688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5796705641375779, "frac_reward_zero_std": 0.6875, "grad_norm": 31.590705883442613, "learning_rate": 5.042419927132886e-07, "loss": 0.0192, "num_tokens": 1640660126.0, "reward": 0.041015625, "reward_std": 0.1038198471069336, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 1698 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.669921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1782.0, "completions/mean_length": 1576.466796875, "completions/mean_terminated_length": 619.44970703125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5800119484509687, "frac_reward_zero_std": 0.625, "grad_norm": 11.647857467325819, "learning_rate": 5.037086961580012e-07, "loss": 0.0588, "num_tokens": 1641535565.0, "reward": 0.0625, "reward_std": 0.1287938952445984, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.2422981858253479, "step": 1699 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1991.0, "completions/mean_length": 1681.18359375, "completions/mean_terminated_length": 625.1969604492188, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.5803533327643595, "frac_reward_zero_std": 0.75, "grad_norm": 23.58696955256426, "learning_rate": 5.031754653047528e-07, "loss": 0.0083, "num_tokens": 1642484939.0, "reward": 0.060546875, "reward_std": 0.10182979702949524, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 1700 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.736328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2027.0, "completions/mean_length": 1675.81640625, "completions/mean_terminated_length": 636.459228515625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5806947170777502, "frac_reward_zero_std": 0.8125, "grad_norm": 53.179475515775344, "learning_rate": 5.02642300910367e-07, "loss": 0.0187, "num_tokens": 1643418653.0, "reward": 0.03515625, "reward_std": 0.06856893002986908, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 1701 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.728515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2020.0, "completions/mean_length": 1684.12109375, "completions/mean_terminated_length": 707.6690673828125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.581036101391141, "frac_reward_zero_std": 0.8125, "grad_norm": 16.73481878408291, "learning_rate": 5.021092037315733e-07, "loss": -0.0108, "num_tokens": 1644351307.0, "reward": 0.04296875, "reward_std": 0.05738259106874466, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 1702 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1896.0, "completions/mean_length": 1707.24609375, "completions/mean_terminated_length": 726.2879028320312, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5813774857045318, "frac_reward_zero_std": 0.8125, "grad_norm": 47.33199262944589, "learning_rate": 5.015761745250055e-07, "loss": 0.0014, "num_tokens": 1645301897.0, "reward": 0.064453125, "reward_std": 0.08130794763565063, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 1703 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.712890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1947.0, "completions/mean_length": 1648.5, "completions/mean_terminated_length": 656.544189453125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5817188700179227, "frac_reward_zero_std": 0.8125, "grad_norm": 13.356393611068537, "learning_rate": 5.01043214047201e-07, "loss": 0.0135, "num_tokens": 1646218857.0, "reward": 0.046875, "reward_std": 0.06546888500452042, "rewards/accuracy_reward/mean": 0.04838709533214569, "rewards/accuracy_reward/std": 0.21479946374893188, "step": 1704 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.748046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1996.0, "completions/mean_length": 1691.810546875, "completions/mean_terminated_length": 634.2868041992188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5820602543313135, "frac_reward_zero_std": 0.65625, "grad_norm": 26.975248150539183, "learning_rate": 5.005103230546e-07, "loss": 0.031, "num_tokens": 1647160648.0, "reward": 0.083984375, "reward_std": 0.13190844655036926, "rewards/accuracy_reward/mean": 0.083984375, "rewards/accuracy_reward/std": 0.2776356339454651, "step": 1705 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.744140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1900.0, "completions/mean_length": 1677.244140625, "completions/mean_terminated_length": 598.9389038085938, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.5824016386447043, "frac_reward_zero_std": 0.875, "grad_norm": 41.0284858154736, "learning_rate": 4.999775023035438e-07, "loss": -0.0001, "num_tokens": 1648099109.0, "reward": 0.052734375, "reward_std": 0.04219770431518555, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 1706 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.685546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1992.0, "completions/mean_length": 1606.44921875, "completions/mean_terminated_length": 643.8136596679688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5827430229580951, "frac_reward_zero_std": 0.6875, "grad_norm": 65.84761392871681, "learning_rate": 4.994447525502735e-07, "loss": 0.0129, "num_tokens": 1649000811.0, "reward": 0.08203125, "reward_std": 0.11043406277894974, "rewards/accuracy_reward/mean": 0.08203125, "rewards/accuracy_reward/std": 0.2746807038784027, "step": 1707 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.626953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2020.0, "completions/mean_length": 1534.900390625, "completions/mean_terminated_length": 672.5706787109375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5830844072714859, "frac_reward_zero_std": 0.65625, "grad_norm": 80.67921401924828, "learning_rate": 4.989120745509305e-07, "loss": 0.0219, "num_tokens": 1649868904.0, "reward": 0.09375, "reward_std": 0.13682594895362854, "rewards/accuracy_reward/mean": 0.09375, "rewards/accuracy_reward/std": 0.29176566004753113, "step": 1708 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2025.0, "completions/mean_length": 1647.4609375, "completions/mean_terminated_length": 583.1714477539062, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5834257915848767, "frac_reward_zero_std": 0.78125, "grad_norm": 15.803462917662138, "learning_rate": 4.983794690615535e-07, "loss": 0.0041, "num_tokens": 1650795796.0, "reward": 0.033203125, "reward_std": 0.07250870764255524, "rewards/accuracy_reward/mean": 0.033203125, "rewards/accuracy_reward/std": 0.17934183776378632, "step": 1709 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2002.0, "completions/mean_length": 1745.796875, "completions/mean_terminated_length": 714.137939453125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5837671758982674, "frac_reward_zero_std": 0.75, "grad_norm": 25.44127424171362, "learning_rate": 4.978469368380787e-07, "loss": 0.0102, "num_tokens": 1651766396.0, "reward": 0.04296875, "reward_std": 0.09028454124927521, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 1710 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1621.0, "completions/mean_length": 1511.326171875, "completions/mean_terminated_length": 554.8912963867188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5841085602116582, "frac_reward_zero_std": 0.8125, "grad_norm": 38.64146044353377, "learning_rate": 4.97314478636338e-07, "loss": 0.0161, "num_tokens": 1652617187.0, "reward": 0.03515625, "reward_std": 0.06888246536254883, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 1711 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2011.0, "completions/mean_length": 1688.845703125, "completions/mean_terminated_length": 771.0069580078125, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.5844499445250491, "frac_reward_zero_std": 0.75, "grad_norm": 49.88085028182138, "learning_rate": 4.967820952120588e-07, "loss": 0.0268, "num_tokens": 1653552484.0, "reward": 0.0703125, "reward_std": 0.10414530336856842, "rewards/accuracy_reward/mean": 0.0703125, "rewards/accuracy_reward/std": 0.25592297315597534, "step": 1712 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.76171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1895.0, "completions/mean_length": 1721.369140625, "completions/mean_terminated_length": 677.6392822265625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5847913288384399, "frac_reward_zero_std": 0.59375, "grad_norm": 73.99758334615439, "learning_rate": 4.962497873208616e-07, "loss": 0.023, "num_tokens": 1654518129.0, "reward": 0.0625, "reward_std": 0.12329617142677307, "rewards/accuracy_reward/mean": 0.06451612710952759, "rewards/accuracy_reward/std": 0.2459181249141693, "step": 1713 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.615234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1736.0, "completions/mean_length": 1485.966796875, "completions/mean_terminated_length": 587.2842407226562, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5851327131518307, "frac_reward_zero_std": 0.625, "grad_norm": 20.456552504913056, "learning_rate": 4.957175557182601e-07, "loss": 0.0491, "num_tokens": 1655357456.0, "reward": 0.123046875, "reward_std": 0.15058745443820953, "rewards/accuracy_reward/mean": 0.123046875, "rewards/accuracy_reward/std": 0.32881227135658264, "step": 1714 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.755859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1678.0, "completions/mean_length": 1692.564453125, "completions/mean_terminated_length": 592.1360473632812, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5854740974652215, "frac_reward_zero_std": 0.9375, "grad_norm": 5.477576678118939, "learning_rate": 4.9518540115966e-07, "loss": 0.0007, "num_tokens": 1656296433.0, "reward": 0.005859375, "reward_std": 0.01848640665411949, "rewards/accuracy_reward/mean": 0.005859375, "rewards/accuracy_reward/std": 0.07639661431312561, "step": 1715 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.697265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1949.0, "completions/mean_length": 1645.04296875, "completions/mean_terminated_length": 716.94189453125, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.5858154817786123, "frac_reward_zero_std": 0.78125, "grad_norm": 25.021323043434897, "learning_rate": 4.946533244003572e-07, "loss": 0.0115, "num_tokens": 1657214279.0, "reward": 0.0703125, "reward_std": 0.08271066844463348, "rewards/accuracy_reward/mean": 0.0703125, "rewards/accuracy_reward/std": 0.25592297315597534, "step": 1716 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.681640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1968.0, "completions/mean_length": 1609.017578125, "completions/mean_terminated_length": 669.1104125976562, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.5861568660920031, "frac_reward_zero_std": 0.71875, "grad_norm": 17.027668599567413, "learning_rate": 4.941213261955374e-07, "loss": 0.0325, "num_tokens": 1658116064.0, "reward": 0.08203125, "reward_std": 0.1186666265130043, "rewards/accuracy_reward/mean": 0.08203125, "rewards/accuracy_reward/std": 0.2746807038784027, "step": 1717 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.76171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2017.0, "completions/mean_length": 1703.955078125, "completions/mean_terminated_length": 604.1392822265625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5864982504053938, "frac_reward_zero_std": 0.78125, "grad_norm": 30.20596131003759, "learning_rate": 4.935894073002749e-07, "loss": 0.0103, "num_tokens": 1659063033.0, "reward": 0.078125, "reward_std": 0.09319131076335907, "rewards/accuracy_reward/mean": 0.078125, "rewards/accuracy_reward/std": 0.26863065361976624, "step": 1718 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.623046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1997.0, "completions/mean_length": 1495.46875, "completions/mean_terminated_length": 582.2175903320312, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5868396347187846, "frac_reward_zero_std": 0.71875, "grad_norm": 15.271407438282715, "learning_rate": 4.930575684695309e-07, "loss": 0.0098, "num_tokens": 1659903209.0, "reward": 0.072265625, "reward_std": 0.09106838703155518, "rewards/accuracy_reward/mean": 0.072265625, "rewards/accuracy_reward/std": 0.2591804563999176, "step": 1719 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.681640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1986.0, "completions/mean_length": 1569.900390625, "completions/mean_terminated_length": 546.2392578125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5871810190321755, "frac_reward_zero_std": 0.75, "grad_norm": 29.110865906675496, "learning_rate": 4.925258104581534e-07, "loss": 0.023, "num_tokens": 1660779878.0, "reward": 0.05859375, "reward_std": 0.08450747281312943, "rewards/accuracy_reward/mean": 0.05859375, "rewards/accuracy_reward/std": 0.23509246110916138, "step": 1720 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.66796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1849.0, "completions/mean_length": 1576.158203125, "completions/mean_terminated_length": 626.9235229492188, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.5875224033455663, "frac_reward_zero_std": 0.75, "grad_norm": 67.27003016349549, "learning_rate": 4.91994134020876e-07, "loss": -0.0038, "num_tokens": 1661659063.0, "reward": 0.10546875, "reward_std": 0.08188469707965851, "rewards/accuracy_reward/mean": 0.1088709682226181, "rewards/accuracy_reward/std": 0.31179171800613403, "step": 1721 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.712890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.0, "completions/mean_length": 1661.4609375, "completions/mean_terminated_length": 702.0679931640625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5878637876589571, "frac_reward_zero_std": 0.78125, "grad_norm": 72.42316134213137, "learning_rate": 4.914625399123159e-07, "loss": 0.014, "num_tokens": 1662588931.0, "reward": 0.064453125, "reward_std": 0.08864850550889969, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 1722 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.80078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1925.0, "completions/mean_length": 1798.005859375, "completions/mean_terminated_length": 793.1275024414062, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.5882051719723479, "frac_reward_zero_std": 0.75, "grad_norm": 97.4499616600289, "learning_rate": 4.909310288869737e-07, "loss": 0.0019, "num_tokens": 1663596950.0, "reward": 0.0703125, "reward_std": 0.10893170535564423, "rewards/accuracy_reward/mean": 0.0703125, "rewards/accuracy_reward/std": 0.25592297315597534, "step": 1723 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.677734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2024.0, "completions/mean_length": 1603.669921875, "completions/mean_terminated_length": 669.8120727539062, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5885465562857387, "frac_reward_zero_std": 0.84375, "grad_norm": 5.696893799314998, "learning_rate": 4.903996016992323e-07, "loss": 0.0034, "num_tokens": 1664501405.0, "reward": 0.01171875, "reward_std": 0.04192390665411949, "rewards/accuracy_reward/mean": 0.01171875, "rewards/accuracy_reward/std": 0.10772226005792618, "step": 1724 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2021.0, "completions/mean_length": 1629.69921875, "completions/mean_terminated_length": 675.1154174804688, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.5888879405991295, "frac_reward_zero_std": 0.875, "grad_norm": 19.573690330793777, "learning_rate": 4.898682591033551e-07, "loss": 0.0066, "num_tokens": 1665406355.0, "reward": 0.017578125, "reward_std": 0.04125870764255524, "rewards/accuracy_reward/mean": 0.018145160749554634, "rewards/accuracy_reward/std": 0.1336110383272171, "step": 1725 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2003.0, "completions/mean_length": 1646.193359375, "completions/mean_terminated_length": 657.9662475585938, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5892293249125202, "frac_reward_zero_std": 0.65625, "grad_norm": 21.092068912595174, "learning_rate": 4.893370018534858e-07, "loss": 0.0162, "num_tokens": 1666326678.0, "reward": 0.08984375, "reward_std": 0.13041163980960846, "rewards/accuracy_reward/mean": 0.0927419364452362, "rewards/accuracy_reward/std": 0.2903633117675781, "step": 1726 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2005.0, "completions/mean_length": 1659.6484375, "completions/mean_terminated_length": 686.1095581054688, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.589570709225911, "frac_reward_zero_std": 0.78125, "grad_norm": 6.006908596141031, "learning_rate": 4.888058307036468e-07, "loss": 0.0086, "num_tokens": 1667256066.0, "reward": 0.03125, "reward_std": 0.0677972063422203, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 1727 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.63671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1750.0, "completions/mean_length": 1550.193359375, "completions/mean_terminated_length": 677.6935424804688, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5899120935393019, "frac_reward_zero_std": 0.75, "grad_norm": 19.056446426509225, "learning_rate": 4.882747464077388e-07, "loss": 0.013, "num_tokens": 1668118901.0, "reward": 0.06640625, "reward_std": 0.09739763289690018, "rewards/accuracy_reward/mean": 0.06854838877916336, "rewards/accuracy_reward/std": 0.25293970108032227, "step": 1728 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2008.0, "completions/mean_length": 1719.982421875, "completions/mean_terminated_length": 694.0564575195312, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.5902534778526927, "frac_reward_zero_std": 0.78125, "grad_norm": 19.029504921133565, "learning_rate": 4.877437497195385e-07, "loss": 0.0122, "num_tokens": 1669088524.0, "reward": 0.0703125, "reward_std": 0.0944955125451088, "rewards/accuracy_reward/mean": 0.07258064299821854, "rewards/accuracy_reward/std": 0.25970885157585144, "step": 1729 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.67578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1899.0, "completions/mean_length": 1620.0546875, "completions/mean_terminated_length": 728.072265625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5905948621660835, "frac_reward_zero_std": 0.84375, "grad_norm": 14.60317749699828, "learning_rate": 4.872128413926989e-07, "loss": 0.0284, "num_tokens": 1669996504.0, "reward": 0.048828125, "reward_std": 0.0681830644607544, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 1730 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1981.0, "completions/mean_length": 1702.923828125, "completions/mean_terminated_length": 709.5227661132812, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.5909362464794743, "frac_reward_zero_std": 0.78125, "grad_norm": 17.68389753572875, "learning_rate": 4.866820221807467e-07, "loss": 0.0227, "num_tokens": 1670944145.0, "reward": 0.044921875, "reward_std": 0.07603531330823898, "rewards/accuracy_reward/mean": 0.0463709682226181, "rewards/accuracy_reward/std": 0.21049949526786804, "step": 1731 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2014.0, "completions/mean_length": 1599.501953125, "completions/mean_terminated_length": 772.272216796875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.5912776307928651, "frac_reward_zero_std": 0.6875, "grad_norm": 40.621538157366196, "learning_rate": 4.861512928370831e-07, "loss": 0.0153, "num_tokens": 1671842850.0, "reward": 0.087890625, "reward_std": 0.1111922413110733, "rewards/accuracy_reward/mean": 0.087890625, "rewards/accuracy_reward/std": 0.2834126651287079, "step": 1732 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.689453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1794.0, "completions/mean_length": 1594.494140625, "completions/mean_terminated_length": 587.654052734375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5916190151062559, "frac_reward_zero_std": 0.8125, "grad_norm": 23.679976144048887, "learning_rate": 4.856206541149812e-07, "loss": 0.0267, "num_tokens": 1672734639.0, "reward": 0.046875, "reward_std": 0.07509076595306396, "rewards/accuracy_reward/mean": 0.04838709533214569, "rewards/accuracy_reward/std": 0.2147994488477707, "step": 1733 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.669921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1960.0, "completions/mean_length": 1581.107421875, "completions/mean_terminated_length": 633.5089111328125, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.5919603994196466, "frac_reward_zero_std": 0.8125, "grad_norm": 11.622606752148648, "learning_rate": 4.850901067675855e-07, "loss": -0.0064, "num_tokens": 1673623270.0, "reward": 0.06640625, "reward_std": 0.07477159798145294, "rewards/accuracy_reward/mean": 0.06640625, "rewards/accuracy_reward/std": 0.2492343932390213, "step": 1734 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.662109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1891.0, "completions/mean_length": 1578.59765625, "completions/mean_terminated_length": 658.7861328125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5923017837330374, "frac_reward_zero_std": 0.84375, "grad_norm": 14.530176875759707, "learning_rate": 4.845596515479113e-07, "loss": 0.0081, "num_tokens": 1674502760.0, "reward": 0.013671875, "reward_std": 0.04478531330823898, "rewards/accuracy_reward/mean": 0.013671875, "rewards/accuracy_reward/std": 0.1162383034825325, "step": 1735 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.69140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2007.0, "completions/mean_length": 1658.45703125, "completions/mean_terminated_length": 785.68359375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.5926431680464282, "frac_reward_zero_std": 0.71875, "grad_norm": 13.508141020196877, "learning_rate": 4.840292892088423e-07, "loss": -0.0045, "num_tokens": 1675425826.0, "reward": 0.04296875, "reward_std": 0.09671792387962341, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 1736 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.623046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1734.0, "completions/mean_length": 1516.37109375, "completions/mean_terminated_length": 637.6683959960938, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.5929845523598191, "frac_reward_zero_std": 0.65625, "grad_norm": 12.035458062105782, "learning_rate": 4.834990205031314e-07, "loss": 0.035, "num_tokens": 1676272624.0, "reward": 0.0625, "reward_std": 0.12199270725250244, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.2422981858253479, "step": 1737 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.611328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1995.0, "completions/mean_length": 1506.310546875, "completions/mean_terminated_length": 654.3065185546875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5933259366732099, "frac_reward_zero_std": 0.75, "grad_norm": 1.3646619089911172, "learning_rate": 4.829688461833975e-07, "loss": 0.014, "num_tokens": 1677118623.0, "reward": 0.0390625, "reward_std": 0.08752824366092682, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 1738 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.728515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2031.0, "completions/mean_length": 1686.89453125, "completions/mean_terminated_length": 717.8848876953125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5936673209866007, "frac_reward_zero_std": 0.6875, "grad_norm": 23.749728402062377, "learning_rate": 4.824387670021263e-07, "loss": 0.001, "num_tokens": 1678056921.0, "reward": 0.0546875, "reward_std": 0.11308281868696213, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 1739 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.646484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1932.0, "completions/mean_length": 1551.169921875, "completions/mean_terminated_length": 642.6022338867188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5940087052999915, "frac_reward_zero_std": 0.6875, "grad_norm": 60.574503881013854, "learning_rate": 4.819087837116682e-07, "loss": 0.0181, "num_tokens": 1678929120.0, "reward": 0.087890625, "reward_std": 0.11090600490570068, "rewards/accuracy_reward/mean": 0.09072580933570862, "rewards/accuracy_reward/std": 0.2875087857246399, "step": 1740 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.591796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2012.0, "completions/mean_length": 1486.51953125, "completions/mean_terminated_length": 672.5071411132812, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5943500896133823, "frac_reward_zero_std": 0.53125, "grad_norm": 23.30272542789789, "learning_rate": 4.813788970642373e-07, "loss": 0.0205, "num_tokens": 1679767066.0, "reward": 0.10546875, "reward_std": 0.16966770589351654, "rewards/accuracy_reward/mean": 0.10546875, "rewards/accuracy_reward/std": 0.3074568510055542, "step": 1741 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.68359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1865.0, "completions/mean_length": 1629.72265625, "completions/mean_terminated_length": 726.0370483398438, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.594691473926773, "frac_reward_zero_std": 0.78125, "grad_norm": 30.568699189168832, "learning_rate": 4.808491078119114e-07, "loss": 0.0068, "num_tokens": 1680680572.0, "reward": 0.029296875, "reward_std": 0.0723423957824707, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 1742 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.623046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1930.0, "completions/mean_length": 1504.654296875, "completions/mean_terminated_length": 606.58544921875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.5950328582401638, "frac_reward_zero_std": 0.65625, "grad_norm": 27.284667207552566, "learning_rate": 4.803194167066292e-07, "loss": 0.0157, "num_tokens": 1681517835.0, "reward": 0.12109375, "reward_std": 0.14497211575508118, "rewards/accuracy_reward/mean": 0.12109375, "rewards/accuracy_reward/std": 0.3265552520751953, "step": 1743 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.63671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2025.0, "completions/mean_length": 1529.216796875, "completions/mean_terminated_length": 620.2150268554688, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.5953742425535546, "frac_reward_zero_std": 0.84375, "grad_norm": 51.263658397355904, "learning_rate": 4.7978982450019e-07, "loss": 0.0112, "num_tokens": 1682377770.0, "reward": 0.01953125, "reward_std": 0.046875, "rewards/accuracy_reward/mean": 0.01953125, "rewards/accuracy_reward/std": 0.1385180652141571, "step": 1744 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.638671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2031.0, "completions/mean_length": 1558.05078125, "completions/mean_terminated_length": 692.3135375976562, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.5957156268669455, "frac_reward_zero_std": 0.6875, "grad_norm": 39.81239885443345, "learning_rate": 4.792603319442533e-07, "loss": 0.0048, "num_tokens": 1683255380.0, "reward": 0.064453125, "reward_std": 0.11833953857421875, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 1745 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1983.0, "completions/mean_length": 1676.74609375, "completions/mean_terminated_length": 797.4605102539062, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5960570111803363, "frac_reward_zero_std": 0.84375, "grad_norm": 2.8214386899226493, "learning_rate": 4.78730939790337e-07, "loss": 0.0092, "num_tokens": 1684189986.0, "reward": 0.044921875, "reward_std": 0.0628596618771553, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 1746 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.69921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1997.0, "completions/mean_length": 1638.396484375, "completions/mean_terminated_length": 686.2012939453125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5963983954937271, "frac_reward_zero_std": 0.71875, "grad_norm": 16.821216602725464, "learning_rate": 4.782016487898163e-07, "loss": 0.0244, "num_tokens": 1685108109.0, "reward": 0.068359375, "reward_std": 0.11218452453613281, "rewards/accuracy_reward/mean": 0.07056451588869095, "rewards/accuracy_reward/std": 0.25635457038879395, "step": 1747 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.693359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1851.0, "completions/mean_length": 1623.1171875, "completions/mean_terminated_length": 662.3948974609375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5967397798071179, "frac_reward_zero_std": 0.71875, "grad_norm": 25.722893442582166, "learning_rate": 4.77672459693923e-07, "loss": 0.0186, "num_tokens": 1686015097.0, "reward": 0.06640625, "reward_std": 0.10475093126296997, "rewards/accuracy_reward/mean": 0.06854838877916336, "rewards/accuracy_reward/std": 0.25293970108032227, "step": 1748 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.662109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1950.0, "completions/mean_length": 1577.7109375, "completions/mean_terminated_length": 656.161865234375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5970811641205087, "frac_reward_zero_std": 0.6875, "grad_norm": 16.360535070185918, "learning_rate": 4.771433732537446e-07, "loss": 0.0521, "num_tokens": 1686896021.0, "reward": 0.03125, "reward_std": 0.09193411469459534, "rewards/accuracy_reward/mean": 0.032258063554763794, "rewards/accuracy_reward/std": 0.17686307430267334, "step": 1749 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.75, "completions/max_length": 2048.0, "completions/max_terminated_length": 2028.0, "completions/mean_length": 1702.796875, "completions/mean_terminated_length": 667.1875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5974225484338994, "frac_reward_zero_std": 0.71875, "grad_norm": 19.35386551251597, "learning_rate": 4.7661439022022186e-07, "loss": 0.0212, "num_tokens": 1687840749.0, "reward": 0.0390625, "reward_std": 0.09005707502365112, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 1750 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.666015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1956.0, "completions/mean_length": 1565.927734375, "completions/mean_terminated_length": 604.6023559570312, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5977639327472902, "frac_reward_zero_std": 0.65625, "grad_norm": 6.463149345825972, "learning_rate": 4.7608551134415e-07, "loss": 0.0169, "num_tokens": 1688715112.0, "reward": 0.08203125, "reward_std": 0.11454009264707565, "rewards/accuracy_reward/mean": 0.08203125, "rewards/accuracy_reward/std": 0.2746807038784027, "step": 1751 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1983.0, "completions/mean_length": 1681.71484375, "completions/mean_terminated_length": 669.0441284179688, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.598105317060681, "frac_reward_zero_std": 0.6875, "grad_norm": 16.468646235492148, "learning_rate": 4.755567373761755e-07, "loss": 0.0345, "num_tokens": 1689647686.0, "reward": 0.103515625, "reward_std": 0.11740150302648544, "rewards/accuracy_reward/mean": 0.103515625, "rewards/accuracy_reward/std": 0.30492907762527466, "step": 1752 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.716796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1951.0, "completions/mean_length": 1646.951171875, "completions/mean_terminated_length": 632.2138061523438, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.5984467013740719, "frac_reward_zero_std": 0.90625, "grad_norm": 1.5852383603687683, "learning_rate": 4.750280690667965e-07, "loss": 0.0039, "num_tokens": 1690572013.0, "reward": 0.01953125, "reward_std": 0.03449726849794388, "rewards/accuracy_reward/mean": 0.01953125, "rewards/accuracy_reward/std": 0.1385180652141571, "step": 1753 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.662109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1997.0, "completions/mean_length": 1584.78515625, "completions/mean_terminated_length": 677.0982666015625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5987880856874627, "frac_reward_zero_std": 0.6875, "grad_norm": 90.1284583937729, "learning_rate": 4.744995071663609e-07, "loss": 0.0294, "num_tokens": 1691454831.0, "reward": 0.09765625, "reward_std": 0.13793784379959106, "rewards/accuracy_reward/mean": 0.09765625, "rewards/accuracy_reward/std": 0.29713961482048035, "step": 1754 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.736328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1800.0, "completions/mean_length": 1675.76953125, "completions/mean_terminated_length": 636.6592407226562, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.5991294700008535, "frac_reward_zero_std": 0.5625, "grad_norm": 19.892408285078744, "learning_rate": 4.7397105242506576e-07, "loss": 0.0681, "num_tokens": 1692395657.0, "reward": 0.12890625, "reward_std": 0.1824507713317871, "rewards/accuracy_reward/mean": 0.12890625, "rewards/accuracy_reward/std": 0.33542385697364807, "step": 1755 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.697265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2032.0, "completions/mean_length": 1622.421875, "completions/mean_terminated_length": 642.2193603515625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5994708543142443, "frac_reward_zero_std": 0.65625, "grad_norm": 28.438715682904697, "learning_rate": 4.734427055929556e-07, "loss": -0.0038, "num_tokens": 1693307121.0, "reward": 0.052734375, "reward_std": 0.11756782233715057, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 1756 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.744140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2020.0, "completions/mean_length": 1700.12109375, "completions/mean_terminated_length": 688.3511352539062, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.5998122386276351, "frac_reward_zero_std": 0.78125, "grad_norm": 34.76343818835714, "learning_rate": 4.729144674199225e-07, "loss": -0.011, "num_tokens": 1694261855.0, "reward": 0.06640625, "reward_std": 0.077679343521595, "rewards/accuracy_reward/mean": 0.06640625, "rewards/accuracy_reward/std": 0.2492343932390213, "step": 1757 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.669921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1977.0, "completions/mean_length": 1578.748046875, "completions/mean_terminated_length": 626.3609619140625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6001536229410258, "frac_reward_zero_std": 0.65625, "grad_norm": 90.20808201620162, "learning_rate": 4.723863386557037e-07, "loss": 0.0228, "num_tokens": 1695148302.0, "reward": 0.041015625, "reward_std": 0.10743913054466248, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 1758 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.736328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1857.0, "completions/mean_length": 1675.423828125, "completions/mean_terminated_length": 634.9703369140625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6004950072544166, "frac_reward_zero_std": 0.8125, "grad_norm": 16.024118039907094, "learning_rate": 4.7185832004988133e-07, "loss": 0.0116, "num_tokens": 1696075559.0, "reward": 0.060546875, "reward_std": 0.06733252108097076, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 1759 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.693359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2008.0, "completions/mean_length": 1607.8671875, "completions/mean_terminated_length": 612.6624145507812, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6008363915678074, "frac_reward_zero_std": 0.65625, "grad_norm": 23.10944354813371, "learning_rate": 4.713304123518814e-07, "loss": 0.0247, "num_tokens": 1696979219.0, "reward": 0.083984375, "reward_std": 0.13180816173553467, "rewards/accuracy_reward/mean": 0.083984375, "rewards/accuracy_reward/std": 0.2776356339454651, "step": 1760 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.60546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1885.0, "completions/mean_length": 1509.361328125, "completions/mean_terminated_length": 682.7376098632812, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6011777758811983, "frac_reward_zero_std": 0.71875, "grad_norm": 32.272756881966046, "learning_rate": 4.708026163109725e-07, "loss": 0.0028, "num_tokens": 1697827932.0, "reward": 0.060546875, "reward_std": 0.0994330644607544, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 1761 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.748046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2040.0, "completions/mean_length": 1713.361328125, "completions/mean_terminated_length": 719.8217163085938, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6015191601945891, "frac_reward_zero_std": 0.8125, "grad_norm": 10.465911730942768, "learning_rate": 4.7027493267626405e-07, "loss": 0.0048, "num_tokens": 1698780165.0, "reward": 0.048828125, "reward_std": 0.06425705552101135, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 1762 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.642578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1679.0, "completions/mean_length": 1530.82421875, "completions/mean_terminated_length": 601.0382080078125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.6018605445079799, "frac_reward_zero_std": 0.6875, "grad_norm": 56.417287247243806, "learning_rate": 4.69747362196707e-07, "loss": 0.0365, "num_tokens": 1699634619.0, "reward": 0.07421875, "reward_std": 0.12517979741096497, "rewards/accuracy_reward/mean": 0.07661290466785431, "rewards/accuracy_reward/std": 0.2662447690963745, "step": 1763 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.654296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2026.0, "completions/mean_length": 1579.13671875, "completions/mean_terminated_length": 691.7401123046875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6022019288213707, "frac_reward_zero_std": 0.6875, "grad_norm": 12.507491630024033, "learning_rate": 4.692199056210907e-07, "loss": 0.0246, "num_tokens": 1700524865.0, "reward": 0.0859375, "reward_std": 0.12774282693862915, "rewards/accuracy_reward/mean": 0.0859375, "rewards/accuracy_reward/std": 0.28054583072662354, "step": 1764 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.673828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2011.0, "completions/mean_length": 1574.771484375, "completions/mean_terminated_length": 597.1437377929688, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.6025433131347615, "frac_reward_zero_std": 0.84375, "grad_norm": 7.428112265761851, "learning_rate": 4.6869256369804353e-07, "loss": 0.0144, "num_tokens": 1701408076.0, "reward": 0.04296875, "reward_std": 0.06233368441462517, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 1765 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.70703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1962.0, "completions/mean_length": 1626.927734375, "completions/mean_terminated_length": 610.739990234375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6028846974481522, "frac_reward_zero_std": 0.65625, "grad_norm": 64.10763037630339, "learning_rate": 4.6816533717603093e-07, "loss": 0.018, "num_tokens": 1702315735.0, "reward": 0.064453125, "reward_std": 0.11804604530334473, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 1766 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.75390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1995.0, "completions/mean_length": 1748.69921875, "completions/mean_terminated_length": 831.793701171875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.603226081761543, "frac_reward_zero_std": 0.71875, "grad_norm": 24.72551911186027, "learning_rate": 4.676382268033544e-07, "loss": 0.0061, "num_tokens": 1703289853.0, "reward": 0.0546875, "reward_std": 0.09797609597444534, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 1767 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.669921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1889.0, "completions/mean_length": 1561.298828125, "completions/mean_terminated_length": 573.7692260742188, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.6035674660749338, "frac_reward_zero_std": 0.71875, "grad_norm": 58.65711889295814, "learning_rate": 4.671112333281508e-07, "loss": 0.0119, "num_tokens": 1704167302.0, "reward": 0.0546875, "reward_std": 0.10150270164012909, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 1768 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.650390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2044.0, "completions/mean_length": 1535.90625, "completions/mean_terminated_length": 583.240234375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6039088503883246, "frac_reward_zero_std": 0.84375, "grad_norm": 38.48290516411763, "learning_rate": 4.6658435749839087e-07, "loss": 0.019, "num_tokens": 1705025670.0, "reward": 0.0703125, "reward_std": 0.07104447484016418, "rewards/accuracy_reward/mean": 0.07258064299821854, "rewards/accuracy_reward/std": 0.25970885157585144, "step": 1769 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2026.0, "completions/mean_length": 1620.8984375, "completions/mean_terminated_length": 570.45947265625, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.6042502347017155, "frac_reward_zero_std": 0.84375, "grad_norm": 8.91925749806277, "learning_rate": 4.6605760006187857e-07, "loss": 0.0058, "num_tokens": 1705928674.0, "reward": 0.029296875, "reward_std": 0.05628519132733345, "rewards/accuracy_reward/mean": 0.030241934582591057, "rewards/accuracy_reward/std": 0.1714252382516861, "step": 1770 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.712890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2012.0, "completions/mean_length": 1630.6875, "completions/mean_terminated_length": 594.9115600585938, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6045916190151063, "frac_reward_zero_std": 0.75, "grad_norm": 10.939123723452132, "learning_rate": 4.655309617662496e-07, "loss": 0.0048, "num_tokens": 1706845346.0, "reward": 0.03515625, "reward_std": 0.0794091522693634, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 1771 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.673828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1904.0, "completions/mean_length": 1566.890625, "completions/mean_terminated_length": 572.9820556640625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6049330033284971, "frac_reward_zero_std": 0.5625, "grad_norm": 25.378032399529374, "learning_rate": 4.6500444335897094e-07, "loss": 0.0233, "num_tokens": 1707723786.0, "reward": 0.06640625, "reward_std": 0.15393419563770294, "rewards/accuracy_reward/mean": 0.06640625, "rewards/accuracy_reward/std": 0.2492343932390213, "step": 1772 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.666015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2040.0, "completions/mean_length": 1585.423828125, "completions/mean_terminated_length": 662.9766235351562, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6052743876418879, "frac_reward_zero_std": 0.71875, "grad_norm": 16.59547981771342, "learning_rate": 4.6447804558733894e-07, "loss": -0.0009, "num_tokens": 1708616275.0, "reward": 0.046875, "reward_std": 0.09723131358623505, "rewards/accuracy_reward/mean": 0.046875, "rewards/accuracy_reward/std": 0.21157780289649963, "step": 1773 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.701171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1874.0, "completions/mean_length": 1644.56640625, "completions/mean_terminated_length": 697.9476928710938, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6056157719552786, "frac_reward_zero_std": 0.5625, "grad_norm": 56.24683300225862, "learning_rate": 4.6395176919847923e-07, "loss": 0.0259, "num_tokens": 1709536645.0, "reward": 0.0390625, "reward_std": 0.12560540437698364, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 1774 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.751953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1938.0, "completions/mean_length": 1673.552734375, "completions/mean_terminated_length": 538.4172973632812, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6059571562686694, "frac_reward_zero_std": 0.78125, "grad_norm": 26.316897250435495, "learning_rate": 4.634256149393445e-07, "loss": 0.027, "num_tokens": 1710473600.0, "reward": 0.021484375, "reward_std": 0.06519509106874466, "rewards/accuracy_reward/mean": 0.02217741869390011, "rewards/accuracy_reward/std": 0.14740893244743347, "step": 1775 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1895.0, "completions/mean_length": 1596.939453125, "completions/mean_terminated_length": 528.63818359375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6062985405820602, "frac_reward_zero_std": 0.84375, "grad_norm": 3.192807479981179, "learning_rate": 4.6289958355671475e-07, "loss": 0.0163, "num_tokens": 1711368401.0, "reward": 0.013671875, "reward_std": 0.043847277760505676, "rewards/accuracy_reward/mean": 0.013671875, "rewards/accuracy_reward/std": 0.1162383034825325, "step": 1776 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.779296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1849.0, "completions/mean_length": 1754.98828125, "completions/mean_terminated_length": 720.3717041015625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.606639924895451, "frac_reward_zero_std": 0.84375, "grad_norm": 6.693809255990296, "learning_rate": 4.6237367579719535e-07, "loss": 0.0245, "num_tokens": 1712339067.0, "reward": 0.025390625, "reward_std": 0.05452118441462517, "rewards/accuracy_reward/mean": 0.026209676638245583, "rewards/accuracy_reward/std": 0.1599196493625641, "step": 1777 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.689453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1951.0, "completions/mean_length": 1576.892578125, "completions/mean_terminated_length": 530.974853515625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6069813092088419, "frac_reward_zero_std": 0.8125, "grad_norm": 42.58992168796201, "learning_rate": 4.618478924072159e-07, "loss": 0.0121, "num_tokens": 1713222884.0, "reward": 0.060546875, "reward_std": 0.07014855742454529, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 1778 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.74609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1719.0, "completions/mean_length": 1682.626953125, "completions/mean_terminated_length": 608.9923095703125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6073226935222327, "frac_reward_zero_std": 0.75, "grad_norm": 14.794848268205975, "learning_rate": 4.6132223413303e-07, "loss": -0.0017, "num_tokens": 1714158869.0, "reward": 0.037109375, "reward_std": 0.07834454625844955, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 1779 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.73828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1962.0, "completions/mean_length": 1677.1953125, "completions/mean_terminated_length": 631.1940307617188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6076640778356235, "frac_reward_zero_std": 0.8125, "grad_norm": 8.92263605217602, "learning_rate": 4.607967017207135e-07, "loss": 0.0078, "num_tokens": 1715099913.0, "reward": 0.021484375, "reward_std": 0.06024399772286415, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 1780 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1771.0, "completions/mean_length": 1705.587890625, "completions/mean_terminated_length": 537.1293334960938, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6080054621490143, "frac_reward_zero_std": 0.84375, "grad_norm": 8.83278625575297, "learning_rate": 4.602712959161633e-07, "loss": -0.0003, "num_tokens": 1716048758.0, "reward": 0.04296875, "reward_std": 0.050948236137628555, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 1781 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.763671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1944.0, "completions/mean_length": 1706.197265625, "completions/mean_terminated_length": 601.6942138671875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.608346846462405, "frac_reward_zero_std": 0.75, "grad_norm": 79.287740337927, "learning_rate": 4.5974601746509687e-07, "loss": 0.0118, "num_tokens": 1716992907.0, "reward": 0.064453125, "reward_std": 0.0975041538476944, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 1782 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.736328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1900.0, "completions/mean_length": 1664.3046875, "completions/mean_terminated_length": 592.7999877929688, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6086882307757958, "frac_reward_zero_std": 0.75, "grad_norm": 43.95360653340727, "learning_rate": 4.592208671130511e-07, "loss": 0.0079, "num_tokens": 1717923111.0, "reward": 0.064453125, "reward_std": 0.08258409798145294, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 1783 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.78125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1904.0, "completions/mean_length": 1740.263671875, "completions/mean_terminated_length": 641.2053833007812, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6090296150891866, "frac_reward_zero_std": 0.875, "grad_norm": 0.22052354947723599, "learning_rate": 4.5869584560538065e-07, "loss": 0.0169, "num_tokens": 1718901134.0, "reward": 0.03515625, "reward_std": 0.04505910724401474, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 1784 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.748046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1971.0, "completions/mean_length": 1669.23046875, "completions/mean_terminated_length": 544.6666870117188, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6093709994025774, "frac_reward_zero_std": 0.71875, "grad_norm": 53.25865397701371, "learning_rate": 4.5817095368725754e-07, "loss": 0.0266, "num_tokens": 1719833444.0, "reward": 0.078125, "reward_std": 0.10658828169107437, "rewards/accuracy_reward/mean": 0.078125, "rewards/accuracy_reward/std": 0.26863065361976624, "step": 1785 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1820.0, "completions/mean_length": 1720.30078125, "completions/mean_terminated_length": 601.6034545898438, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6097123837159683, "frac_reward_zero_std": 0.875, "grad_norm": 15.336664911287983, "learning_rate": 4.576461921036702e-07, "loss": 0.0072, "num_tokens": 1720792910.0, "reward": 0.009765625, "reward_std": 0.03411140665411949, "rewards/accuracy_reward/mean": 0.009765625, "rewards/accuracy_reward/std": 0.09843364357948303, "step": 1786 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.74609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1849.0, "completions/mean_length": 1674.361328125, "completions/mean_terminated_length": 576.4384765625, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.6100537680293591, "frac_reward_zero_std": 0.84375, "grad_norm": 31.4563161853599, "learning_rate": 4.571215615994212e-07, "loss": 0.0189, "num_tokens": 1721730679.0, "reward": 0.060546875, "reward_std": 0.06750432401895523, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 1787 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1858.0, "completions/mean_length": 1621.626953125, "completions/mean_terminated_length": 552.7739868164062, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6103951523427499, "frac_reward_zero_std": 0.59375, "grad_norm": 165.11636596109886, "learning_rate": 4.565970629191278e-07, "loss": 0.0045, "num_tokens": 1722639448.0, "reward": 0.080078125, "reward_std": 0.13820409774780273, "rewards/accuracy_reward/mean": 0.08266129344701767, "rewards/accuracy_reward/std": 0.2756475806236267, "step": 1788 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1919.0, "completions/mean_length": 1659.427734375, "completions/mean_terminated_length": 626.9357299804688, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.6107365366561407, "frac_reward_zero_std": 0.90625, "grad_norm": 14.591164161902098, "learning_rate": 4.5607269680721993e-07, "loss": 0.0163, "num_tokens": 1723565251.0, "reward": 0.037109375, "reward_std": 0.04461899772286415, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 1789 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.77734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1742.0, "completions/mean_length": 1750.087890625, "completions/mean_terminated_length": 710.0087890625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6110779209695314, "frac_reward_zero_std": 0.84375, "grad_norm": 15.830102008838816, "learning_rate": 4.5554846400793946e-07, "loss": -0.002, "num_tokens": 1724545328.0, "reward": 0.052734375, "reward_std": 0.07499225437641144, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 1790 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1899.0, "completions/mean_length": 1629.111328125, "completions/mean_terminated_length": 579.0205688476562, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.6114193052829222, "frac_reward_zero_std": 0.78125, "grad_norm": 8.346945159707694, "learning_rate": 4.5502436526533896e-07, "loss": 0.0215, "num_tokens": 1725455497.0, "reward": 0.052734375, "reward_std": 0.08318498730659485, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 1791 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.685546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1943.0, "completions/mean_length": 1619.904296875, "completions/mean_terminated_length": 686.6024780273438, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.611760689596313, "frac_reward_zero_std": 0.75, "grad_norm": 22.76036886807291, "learning_rate": 4.5450040132328074e-07, "loss": 0.017, "num_tokens": 1726353624.0, "reward": 0.080078125, "reward_std": 0.10925643146038055, "rewards/accuracy_reward/mean": 0.080078125, "rewards/accuracy_reward/std": 0.271679550409317, "step": 1792 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.70703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2042.0, "completions/mean_length": 1657.447265625, "completions/mean_terminated_length": 715.260009765625, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.6121020739097038, "frac_reward_zero_std": 0.75, "grad_norm": 12.811707480112897, "learning_rate": 4.539765729254356e-07, "loss": 0.0018, "num_tokens": 1727281341.0, "reward": 0.0390625, "reward_std": 0.08698301017284393, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 1793 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.75, "completions/max_length": 2048.0, "completions/max_terminated_length": 2018.0, "completions/mean_length": 1694.591796875, "completions/mean_terminated_length": 634.3671875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.6124434582230946, "frac_reward_zero_std": 0.75, "grad_norm": 19.720413781595724, "learning_rate": 4.5345288081528223e-07, "loss": 0.0218, "num_tokens": 1728229596.0, "reward": 0.046875, "reward_std": 0.09225328266620636, "rewards/accuracy_reward/mean": 0.046875, "rewards/accuracy_reward/std": 0.21157780289649963, "step": 1794 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.67578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1989.0, "completions/mean_length": 1570.12109375, "completions/mean_terminated_length": 574.0602416992188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6127848425364855, "frac_reward_zero_std": 0.625, "grad_norm": 19.631576705001798, "learning_rate": 4.529293257361059e-07, "loss": 0.0234, "num_tokens": 1729122106.0, "reward": 0.11328125, "reward_std": 0.14618882536888123, "rewards/accuracy_reward/mean": 0.11328125, "rewards/accuracy_reward/std": 0.3172462284564972, "step": 1795 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.716796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1931.0, "completions/mean_length": 1660.6171875, "completions/mean_terminated_length": 680.137939453125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6131262268498763, "frac_reward_zero_std": 0.71875, "grad_norm": 35.70968879959113, "learning_rate": 4.5240590843099725e-07, "loss": -0.009, "num_tokens": 1730054486.0, "reward": 0.04296875, "reward_std": 0.09743183106184006, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 1796 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1996.0, "completions/mean_length": 1576.140625, "completions/mean_terminated_length": 643.3953247070312, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6134676111632671, "frac_reward_zero_std": 0.875, "grad_norm": 50.68452388034922, "learning_rate": 4.5188262964285126e-07, "loss": -0.0022, "num_tokens": 1730938686.0, "reward": 0.03125, "reward_std": 0.047646716237068176, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 1797 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2016.0, "completions/mean_length": 1687.685546875, "completions/mean_terminated_length": 650.4166870117188, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.6138089954766578, "frac_reward_zero_std": 0.8125, "grad_norm": 6.231496164555943, "learning_rate": 4.513594901143668e-07, "loss": 0.0109, "num_tokens": 1731893613.0, "reward": 0.015625, "reward_std": 0.05259781330823898, "rewards/accuracy_reward/mean": 0.015625, "rewards/accuracy_reward/std": 0.12414088100194931, "step": 1798 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.78125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1887.0, "completions/mean_length": 1735.412109375, "completions/mean_terminated_length": 619.0267944335938, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.6141503797900486, "frac_reward_zero_std": 0.78125, "grad_norm": 15.186711260204905, "learning_rate": 4.508364905880444e-07, "loss": 0.0093, "num_tokens": 1732858896.0, "reward": 0.029296875, "reward_std": 0.07570268213748932, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 1799 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.666015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2026.0, "completions/mean_length": 1580.34765625, "completions/mean_terminated_length": 647.7777709960938, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.6144917641034394, "frac_reward_zero_std": 0.84375, "grad_norm": 0.8823079623799638, "learning_rate": 4.503136318061863e-07, "loss": -0.0047, "num_tokens": 1733748946.0, "reward": 0.037109375, "reward_std": 0.06568294763565063, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 1800 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.61328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1800.0, "completions/mean_length": 1478.064453125, "completions/mean_terminated_length": 574.227294921875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6148331484168302, "frac_reward_zero_std": 0.71875, "grad_norm": 3.440715382871405, "learning_rate": 4.497909145108949e-07, "loss": 0.0266, "num_tokens": 1734574963.0, "reward": 0.060546875, "reward_std": 0.10618237406015396, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 1801 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.73828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1981.0, "completions/mean_length": 1663.755859375, "completions/mean_terminated_length": 579.84326171875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.615174532730221, "frac_reward_zero_std": 0.65625, "grad_norm": 57.573321466204106, "learning_rate": 4.4926833944407207e-07, "loss": 0.0155, "num_tokens": 1735500566.0, "reward": 0.107421875, "reward_std": 0.1546606868505478, "rewards/accuracy_reward/mean": 0.107421875, "rewards/accuracy_reward/std": 0.30995169281959534, "step": 1802 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.677734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1951.0, "completions/mean_length": 1637.591796875, "completions/mean_terminated_length": 774.4909057617188, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6155159170436119, "frac_reward_zero_std": 0.75, "grad_norm": 14.267590194011289, "learning_rate": 4.4874590734741715e-07, "loss": 0.0149, "num_tokens": 1736409557.0, "reward": 0.04296875, "reward_std": 0.08126020431518555, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 1803 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.728515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2023.0, "completions/mean_length": 1663.822265625, "completions/mean_terminated_length": 632.8992919921875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6158573013570027, "frac_reward_zero_std": 0.78125, "grad_norm": 14.425682702275317, "learning_rate": 4.4822361896242734e-07, "loss": 0.0178, "num_tokens": 1737344122.0, "reward": 0.072265625, "reward_std": 0.07823152095079422, "rewards/accuracy_reward/mean": 0.072265625, "rewards/accuracy_reward/std": 0.2591804563999176, "step": 1804 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.634765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2022.0, "completions/mean_length": 1511.30859375, "completions/mean_terminated_length": 578.55615234375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6161986856703935, "frac_reward_zero_std": 0.65625, "grad_norm": 19.422064369614954, "learning_rate": 4.477014750303951e-07, "loss": -0.0012, "num_tokens": 1738191320.0, "reward": 0.041015625, "reward_std": 0.10689391195774078, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 1805 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.669921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2037.0, "completions/mean_length": 1557.4296875, "completions/mean_terminated_length": 561.775146484375, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.6165400699837843, "frac_reward_zero_std": 0.75, "grad_norm": 14.784035150807018, "learning_rate": 4.471794762924084e-07, "loss": 0.0152, "num_tokens": 1739070228.0, "reward": 0.060546875, "reward_std": 0.0815330445766449, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 1806 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.65234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1954.0, "completions/mean_length": 1532.873046875, "completions/mean_terminated_length": 566.5786743164062, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.616881454297175, "frac_reward_zero_std": 0.65625, "grad_norm": 58.40928013189856, "learning_rate": 4.46657623489349e-07, "loss": 0.0341, "num_tokens": 1739939779.0, "reward": 0.091796875, "reward_std": 0.13667413592338562, "rewards/accuracy_reward/mean": 0.091796875, "rewards/accuracy_reward/std": 0.289021372795105, "step": 1807 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.66796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1956.0, "completions/mean_length": 1577.46875, "completions/mean_terminated_length": 630.87060546875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6172228386105658, "frac_reward_zero_std": 0.6875, "grad_norm": 21.973966482179144, "learning_rate": 4.461359173618914e-07, "loss": 0.0056, "num_tokens": 1740830259.0, "reward": 0.060546875, "reward_std": 0.11384069174528122, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 1808 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.630859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1984.0, "completions/mean_length": 1534.32421875, "completions/mean_terminated_length": 656.4550170898438, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6175642229239566, "frac_reward_zero_std": 0.78125, "grad_norm": 24.589240238210678, "learning_rate": 4.456143586505021e-07, "loss": 0.0017, "num_tokens": 1741688201.0, "reward": 0.046875, "reward_std": 0.08078034222126007, "rewards/accuracy_reward/mean": 0.046875, "rewards/accuracy_reward/std": 0.21157780289649963, "step": 1809 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.697265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1928.0, "completions/mean_length": 1596.892578125, "completions/mean_terminated_length": 557.8903198242188, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.6179056072373474, "frac_reward_zero_std": 0.71875, "grad_norm": 31.959621530188844, "learning_rate": 4.450929480954383e-07, "loss": 0.0131, "num_tokens": 1742583650.0, "reward": 0.044921875, "reward_std": 0.09286519140005112, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 1810 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.67578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1933.0, "completions/mean_length": 1594.419921875, "completions/mean_terminated_length": 649.0059814453125, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.6182469915507383, "frac_reward_zero_std": 0.8125, "grad_norm": 21.597695409867235, "learning_rate": 4.445716864367472e-07, "loss": 0.0001, "num_tokens": 1743476953.0, "reward": 0.03125, "reward_std": 0.06837663054466248, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 1811 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1905.0, "completions/mean_length": 1629.60546875, "completions/mean_terminated_length": 580.75341796875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6185883758641291, "frac_reward_zero_std": 0.8125, "grad_norm": 9.561996206174552, "learning_rate": 4.440505744142639e-07, "loss": 0.0104, "num_tokens": 1744382607.0, "reward": 0.0234375, "reward_std": 0.062167368829250336, "rewards/accuracy_reward/mean": 0.0234375, "rewards/accuracy_reward/std": 0.15143637359142303, "step": 1812 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2014.0, "completions/mean_length": 1653.640625, "completions/mean_terminated_length": 665.0410766601562, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.6189297601775199, "frac_reward_zero_std": 0.875, "grad_norm": 4.187602275271422, "learning_rate": 4.4352961276761183e-07, "loss": 0.0111, "num_tokens": 1745301959.0, "reward": 0.029296875, "reward_std": 0.046722229570150375, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 1813 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.732421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1993.0, "completions/mean_length": 1677.26953125, "completions/mean_terminated_length": 662.496337890625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.6192711444909107, "frac_reward_zero_std": 0.75, "grad_norm": 15.76920017639966, "learning_rate": 4.4300880223620063e-07, "loss": 0.0168, "num_tokens": 1746237233.0, "reward": 0.0546875, "reward_std": 0.0868699848651886, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 1814 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.759765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1980.0, "completions/mean_length": 1697.10546875, "completions/mean_terminated_length": 587.3658447265625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6196125288043014, "frac_reward_zero_std": 0.75, "grad_norm": 9.48118460668332, "learning_rate": 4.424881435592256e-07, "loss": 0.0085, "num_tokens": 1747185191.0, "reward": 0.0546875, "reward_std": 0.09518137574195862, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 1815 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.638671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1720.0, "completions/mean_length": 1520.03125, "completions/mean_terminated_length": 586.8108520507812, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.6199539131176922, "frac_reward_zero_std": 0.65625, "grad_norm": 34.61352425813081, "learning_rate": 4.419676374756668e-07, "loss": 0.0332, "num_tokens": 1748037863.0, "reward": 0.1171875, "reward_std": 0.14920367300510406, "rewards/accuracy_reward/mean": 0.1171875, "rewards/accuracy_reward/std": 0.32195815443992615, "step": 1816 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1761.0, "completions/mean_length": 1673.232421875, "completions/mean_terminated_length": 500.9193420410156, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.620295297431083, "frac_reward_zero_std": 0.90625, "grad_norm": 10.577681196747255, "learning_rate": 4.4144728472428725e-07, "loss": 0.0081, "num_tokens": 1748979198.0, "reward": 0.0078125, "reward_std": 0.02629890665411949, "rewards/accuracy_reward/mean": 0.0078125, "rewards/accuracy_reward/std": 0.08812850713729858, "step": 1817 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.701171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1809.0, "completions/mean_length": 1586.27734375, "completions/mean_terminated_length": 502.8888854980469, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6206366817444738, "frac_reward_zero_std": 0.78125, "grad_norm": 26.38431695122219, "learning_rate": 4.409270860436325e-07, "loss": 0.0028, "num_tokens": 1749875692.0, "reward": 0.09765625, "reward_std": 0.09060370922088623, "rewards/accuracy_reward/mean": 0.09765625, "rewards/accuracy_reward/std": 0.29713961482048035, "step": 1818 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.751953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2043.0, "completions/mean_length": 1696.716796875, "completions/mean_terminated_length": 631.8031616210938, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6209780660578647, "frac_reward_zero_std": 0.84375, "grad_norm": 50.101126611449025, "learning_rate": 4.404070421720293e-07, "loss": 0.0046, "num_tokens": 1750826187.0, "reward": 0.029296875, "reward_std": 0.05688370764255524, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 1819 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.68359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1858.0, "completions/mean_length": 1579.259765625, "completions/mean_terminated_length": 566.5493774414062, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6213194503712555, "frac_reward_zero_std": 0.65625, "grad_norm": 25.579078851494202, "learning_rate": 4.398871538475849e-07, "loss": 0.02, "num_tokens": 1751710992.0, "reward": 0.087890625, "reward_std": 0.11630964279174805, "rewards/accuracy_reward/mean": 0.087890625, "rewards/accuracy_reward/std": 0.2834126651287079, "step": 1820 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.740234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1740.0, "completions/mean_length": 1690.267578125, "completions/mean_terminated_length": 670.8646850585938, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6216608346846463, "frac_reward_zero_std": 0.875, "grad_norm": 7.1482583364896275, "learning_rate": 4.3936742180818565e-07, "loss": 0.0069, "num_tokens": 1752662265.0, "reward": 0.03125, "reward_std": 0.047646716237068176, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 1821 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.76953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2040.0, "completions/mean_length": 1703.916015625, "completions/mean_terminated_length": 555.0254516601562, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6220022189980371, "frac_reward_zero_std": 0.8125, "grad_norm": 11.334885314753166, "learning_rate": 4.3884784679149613e-07, "loss": 0.0194, "num_tokens": 1753615118.0, "reward": 0.02734375, "reward_std": 0.0645298957824707, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 1822 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.798828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1940.0, "completions/mean_length": 1748.908203125, "completions/mean_terminated_length": 561.25244140625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.6223436033114278, "frac_reward_zero_std": 0.6875, "grad_norm": 60.002725650760134, "learning_rate": 4.383284295349583e-07, "loss": 0.017, "num_tokens": 1754590031.0, "reward": 0.044921875, "reward_std": 0.10336729884147644, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 1823 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1547.0, "completions/mean_length": 1697.1640625, "completions/mean_terminated_length": 551.1000366210938, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.6226849876248186, "frac_reward_zero_std": 0.6875, "grad_norm": 65.86490336253512, "learning_rate": 4.378091707757896e-07, "loss": 0.0029, "num_tokens": 1755533731.0, "reward": 0.052734375, "reward_std": 0.10145636647939682, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 1824 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.775390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1761.0, "completions/mean_length": 1721.30078125, "completions/mean_terminated_length": 593.4782104492188, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6230263719382094, "frac_reward_zero_std": 0.78125, "grad_norm": 7.726497648827249, "learning_rate": 4.372900712509831e-07, "loss": 0.0321, "num_tokens": 1756485885.0, "reward": 0.041015625, "reward_std": 0.0830162987112999, "rewards/accuracy_reward/mean": 0.04233871027827263, "rewards/accuracy_reward/std": 0.2015640139579773, "step": 1825 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1883.0, "completions/mean_length": 1600.880859375, "completions/mean_terminated_length": 617.21875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6233677562516002, "frac_reward_zero_std": 0.65625, "grad_norm": 21.824288767814878, "learning_rate": 4.367711316973054e-07, "loss": 0.0196, "num_tokens": 1757382000.0, "reward": 0.08984375, "reward_std": 0.13093778491020203, "rewards/accuracy_reward/mean": 0.08984375, "rewards/accuracy_reward/std": 0.2862374484539032, "step": 1826 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.720703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1749.0, "completions/mean_length": 1634.39453125, "completions/mean_terminated_length": 567.5174560546875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.623709140564991, "frac_reward_zero_std": 0.75, "grad_norm": 27.525159576689205, "learning_rate": 4.3625235285129634e-07, "loss": 0.0059, "num_tokens": 1758293562.0, "reward": 0.029296875, "reward_std": 0.08010855317115784, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 1827 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.712890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2026.0, "completions/mean_length": 1642.173828125, "completions/mean_terminated_length": 634.5101928710938, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.6240505248783819, "frac_reward_zero_std": 0.78125, "grad_norm": 10.182693050655088, "learning_rate": 4.3573373544926786e-07, "loss": 0.0114, "num_tokens": 1759215939.0, "reward": 0.052734375, "reward_std": 0.07712717354297638, "rewards/accuracy_reward/mean": 0.05443548411130905, "rewards/accuracy_reward/std": 0.227104052901268, "step": 1828 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.728515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1779.0, "completions/mean_length": 1658.236328125, "completions/mean_terminated_length": 612.32373046875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6243919091917727, "frac_reward_zero_std": 0.9375, "grad_norm": 1.3930119646294938, "learning_rate": 4.352152802273024e-07, "loss": 0.0169, "num_tokens": 1760140572.0, "reward": 0.015625, "reward_std": 0.023823359981179237, "rewards/accuracy_reward/mean": 0.015625, "rewards/accuracy_reward/std": 0.12414088100194931, "step": 1829 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.70703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1690.0, "completions/mean_length": 1610.041015625, "completions/mean_terminated_length": 553.1000366210938, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6247332935051635, "frac_reward_zero_std": 0.75, "grad_norm": 28.483093153318027, "learning_rate": 4.3469698792125196e-07, "loss": 0.0237, "num_tokens": 1761041649.0, "reward": 0.033203125, "reward_std": 0.07686128467321396, "rewards/accuracy_reward/mean": 0.033203125, "rewards/accuracy_reward/std": 0.17934183776378632, "step": 1830 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.744140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1947.0, "completions/mean_length": 1673.59375, "completions/mean_terminated_length": 584.6717529296875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6250746778185542, "frac_reward_zero_std": 0.6875, "grad_norm": 86.36712319459413, "learning_rate": 4.341788592667381e-07, "loss": 0.031, "num_tokens": 1761981665.0, "reward": 0.09765625, "reward_std": 0.12797029316425323, "rewards/accuracy_reward/mean": 0.09765625, "rewards/accuracy_reward/std": 0.29713961482048035, "step": 1831 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.724609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1939.0, "completions/mean_length": 1648.947265625, "completions/mean_terminated_length": 598.9573974609375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.625416062131945, "frac_reward_zero_std": 0.75, "grad_norm": 9.152933770072398, "learning_rate": 4.336608949991496e-07, "loss": 0.0188, "num_tokens": 1762908902.0, "reward": 0.03515625, "reward_std": 0.08493966609239578, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 1832 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.732421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2023.0, "completions/mean_length": 1661.40625, "completions/mean_terminated_length": 603.211669921875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.6257574464453358, "frac_reward_zero_std": 0.71875, "grad_norm": 17.800753345782766, "learning_rate": 4.3314309585364185e-07, "loss": 0.0237, "num_tokens": 1763839094.0, "reward": 0.04296875, "reward_std": 0.09338457882404327, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 1833 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1955.0, "completions/mean_length": 1619.619140625, "completions/mean_terminated_length": 605.0328979492188, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.6260988307587266, "frac_reward_zero_std": 0.78125, "grad_norm": 7.519855647768585, "learning_rate": 4.3262546256513613e-07, "loss": 0.0101, "num_tokens": 1764746163.0, "reward": 0.046875, "reward_std": 0.07898354530334473, "rewards/accuracy_reward/mean": 0.046875, "rewards/accuracy_reward/std": 0.21157780289649963, "step": 1834 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1878.0, "completions/mean_length": 1635.24609375, "completions/mean_terminated_length": 620.0946044921875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6264402150721174, "frac_reward_zero_std": 0.8125, "grad_norm": 119.7640821713971, "learning_rate": 4.3210799586831825e-07, "loss": 0.0052, "num_tokens": 1765660849.0, "reward": 0.0390625, "reward_std": 0.0712449923157692, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 1835 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.751953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1944.0, "completions/mean_length": 1693.01171875, "completions/mean_terminated_length": 616.8661499023438, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6267815993855083, "frac_reward_zero_std": 0.84375, "grad_norm": 10.961395164986127, "learning_rate": 4.3159069649763747e-07, "loss": 0.0076, "num_tokens": 1766613319.0, "reward": 0.02734375, "reward_std": 0.054907046258449554, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 1836 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.689453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1871.0, "completions/mean_length": 1596.236328125, "completions/mean_terminated_length": 593.26416015625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6271229836988991, "frac_reward_zero_std": 0.84375, "grad_norm": 27.529488416793896, "learning_rate": 4.3107356518730564e-07, "loss": 0.0217, "num_tokens": 1767503200.0, "reward": 0.0234375, "reward_std": 0.051659777760505676, "rewards/accuracy_reward/mean": 0.0234375, "rewards/accuracy_reward/std": 0.15143637359142303, "step": 1837 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.755859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1536.0, "completions/mean_length": 1691.791015625, "completions/mean_terminated_length": 588.968017578125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6274643680122899, "frac_reward_zero_std": 0.84375, "grad_norm": 28.010663518769846, "learning_rate": 4.30556602671296e-07, "loss": 0.0028, "num_tokens": 1768444325.0, "reward": 0.03125, "reward_std": 0.05880707502365112, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 1838 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2012.0, "completions/mean_length": 1699.998046875, "completions/mean_terminated_length": 611.0886840820312, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6278057523256806, "frac_reward_zero_std": 0.75, "grad_norm": 5.85948075817655, "learning_rate": 4.300398096833424e-07, "loss": 0.0053, "num_tokens": 1769391332.0, "reward": 0.064453125, "reward_std": 0.09280546009540558, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 1839 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.75, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1697.91796875, "completions/mean_terminated_length": 647.671875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.6281471366390714, "frac_reward_zero_std": 0.75, "grad_norm": 107.37095180902176, "learning_rate": 4.2952318695693803e-07, "loss": -0.0041, "num_tokens": 1770330138.0, "reward": 0.0625, "reward_std": 0.10244170576334, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.2422981858253479, "step": 1840 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.724609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2033.0, "completions/mean_length": 1654.841796875, "completions/mean_terminated_length": 620.723388671875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.6284885209524622, "frac_reward_zero_std": 0.6875, "grad_norm": 28.36705516977205, "learning_rate": 4.290067352253344e-07, "loss": 0.0166, "num_tokens": 1771258457.0, "reward": 0.068359375, "reward_std": 0.10353910177946091, "rewards/accuracy_reward/mean": 0.068359375, "rewards/accuracy_reward/std": 0.25260838866233826, "step": 1841 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.716796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1598.0, "completions/mean_length": 1633.423828125, "completions/mean_terminated_length": 584.1172485351562, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.628829905265853, "frac_reward_zero_std": 0.8125, "grad_norm": 36.13622735992118, "learning_rate": 4.2849045522153994e-07, "loss": 0.0079, "num_tokens": 1772171618.0, "reward": 0.05078125, "reward_std": 0.07807311415672302, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 1842 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1962.0, "completions/mean_length": 1601.009765625, "completions/mean_terminated_length": 776.5611572265625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6291712895792438, "frac_reward_zero_std": 0.75, "grad_norm": 1.2017222626306845, "learning_rate": 4.2797434767831985e-07, "loss": 0.014, "num_tokens": 1773059063.0, "reward": 0.0390625, "reward_std": 0.08177263289690018, "rewards/accuracy_reward/mean": 0.04032257944345474, "rewards/accuracy_reward/std": 0.19691328704357147, "step": 1843 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.693359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1989.0, "completions/mean_length": 1623.66796875, "completions/mean_terminated_length": 664.1911010742188, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.6295126738926347, "frac_reward_zero_std": 0.6875, "grad_norm": 19.6054492603802, "learning_rate": 4.274584133281944e-07, "loss": -0.006, "num_tokens": 1773966365.0, "reward": 0.0625, "reward_std": 0.10904473811388016, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.2422981858253479, "step": 1844 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1974.0, "completions/mean_length": 1489.626953125, "completions/mean_terminated_length": 527.324462890625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.6298540582060255, "frac_reward_zero_std": 0.84375, "grad_norm": 2.5058880491406486, "learning_rate": 4.269426529034382e-07, "loss": 0.0262, "num_tokens": 1774800990.0, "reward": 0.025390625, "reward_std": 0.05303792655467987, "rewards/accuracy_reward/mean": 0.025390625, "rewards/accuracy_reward/std": 0.15746226906776428, "step": 1845 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.64453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1948.0, "completions/mean_length": 1535.265625, "completions/mean_terminated_length": 605.5824584960938, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.6301954425194163, "frac_reward_zero_std": 0.75, "grad_norm": 30.826239809888694, "learning_rate": 4.264270671360787e-07, "loss": 0.0048, "num_tokens": 1775668006.0, "reward": 0.048828125, "reward_std": 0.09231997281312943, "rewards/accuracy_reward/mean": 0.05040322616696358, "rewards/accuracy_reward/std": 0.21899642050266266, "step": 1846 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.685546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1966.0, "completions/mean_length": 1610.08984375, "completions/mean_terminated_length": 655.3912963867188, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.630536826832807, "frac_reward_zero_std": 0.65625, "grad_norm": 117.83752239192344, "learning_rate": 4.2591165675789555e-07, "loss": 0.0386, "num_tokens": 1776564260.0, "reward": 0.0703125, "reward_std": 0.1254933476448059, "rewards/accuracy_reward/mean": 0.0703125, "rewards/accuracy_reward/std": 0.25592297315597534, "step": 1847 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1764.0, "completions/mean_length": 1571.625, "completions/mean_terminated_length": 523.6000366210938, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6308782111461978, "frac_reward_zero_std": 0.8125, "grad_norm": 30.961108684547437, "learning_rate": 4.2539642250041973e-07, "loss": 0.0148, "num_tokens": 1777442372.0, "reward": 0.048828125, "reward_std": 0.07438573986291885, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 1848 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2028.0, "completions/mean_length": 1691.486328125, "completions/mean_terminated_length": 665.1591186523438, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6312195954595886, "frac_reward_zero_std": 0.6875, "grad_norm": 24.969633047847786, "learning_rate": 4.2488136509493165e-07, "loss": 0.006, "num_tokens": 1778387517.0, "reward": 0.064453125, "reward_std": 0.11509227007627487, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 1849 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.716796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1980.0, "completions/mean_length": 1672.365234375, "completions/mean_terminated_length": 722.027587890625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.6315609797729794, "frac_reward_zero_std": 0.78125, "grad_norm": 11.342091751656254, "learning_rate": 4.243664852724614e-07, "loss": 0.011, "num_tokens": 1779339384.0, "reward": 0.0546875, "reward_std": 0.088299959897995, "rewards/accuracy_reward/mean": 0.05645161122083664, "rewards/accuracy_reward/std": 0.23102475702762604, "step": 1850 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2034.0, "completions/mean_length": 1514.173828125, "completions/mean_terminated_length": 681.6699829101562, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6319023640863702, "frac_reward_zero_std": 0.59375, "grad_norm": 45.96276231829346, "learning_rate": 4.238517837637866e-07, "loss": 0.0385, "num_tokens": 1780192625.0, "reward": 0.107421875, "reward_std": 0.16086195409297943, "rewards/accuracy_reward/mean": 0.107421875, "rewards/accuracy_reward/std": 0.30995169281959534, "step": 1851 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.68359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2012.0, "completions/mean_length": 1635.52734375, "completions/mean_terminated_length": 744.3827514648438, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.632243748399761, "frac_reward_zero_std": 0.625, "grad_norm": 10.33648519828817, "learning_rate": 4.233372612994319e-07, "loss": 0.037, "num_tokens": 1781105791.0, "reward": 0.060546875, "reward_std": 0.1282098889350891, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 1852 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.75, "completions/max_length": 2048.0, "completions/max_terminated_length": 1976.0, "completions/mean_length": 1689.939453125, "completions/mean_terminated_length": 615.7578125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.6325851327131519, "frac_reward_zero_std": 0.9375, "grad_norm": 7.116431464790528, "learning_rate": 4.228229186096679e-07, "loss": 0.0028, "num_tokens": 1782061056.0, "reward": 0.013671875, "reward_std": 0.026572702452540398, "rewards/accuracy_reward/mean": 0.013671875, "rewards/accuracy_reward/std": 0.1162383034825325, "step": 1853 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.75390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1917.0, "completions/mean_length": 1740.44140625, "completions/mean_terminated_length": 798.2381591796875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6329265170265427, "frac_reward_zero_std": 0.84375, "grad_norm": 3.756186788779025, "learning_rate": 4.223087564245099e-07, "loss": 0.0174, "num_tokens": 1783026546.0, "reward": 0.021484375, "reward_std": 0.05287161096930504, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 1854 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.66796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1839.0, "completions/mean_length": 1572.73046875, "completions/mean_terminated_length": 616.6000366210938, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6332679013399334, "frac_reward_zero_std": 0.78125, "grad_norm": 18.840068506715816, "learning_rate": 4.2179477547371713e-07, "loss": 0.008, "num_tokens": 1783909000.0, "reward": 0.025390625, "reward_std": 0.06536141037940979, "rewards/accuracy_reward/mean": 0.025390625, "rewards/accuracy_reward/std": 0.15746226906776428, "step": 1855 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.67578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1920.0, "completions/mean_length": 1575.8671875, "completions/mean_terminated_length": 591.7830810546875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6336092856533242, "frac_reward_zero_std": 0.75, "grad_norm": 4.193641523439476, "learning_rate": 4.212809764867915e-07, "loss": 0.0102, "num_tokens": 1784790500.0, "reward": 0.0234375, "reward_std": 0.07300759106874466, "rewards/accuracy_reward/mean": 0.0234375, "rewards/accuracy_reward/std": 0.15143637359142303, "step": 1856 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1933.0, "completions/mean_length": 1642.703125, "completions/mean_terminated_length": 522.1764526367188, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.633950669966715, "frac_reward_zero_std": 0.71875, "grad_norm": 23.448024839000897, "learning_rate": 4.2076736019297674e-07, "loss": 0.0218, "num_tokens": 1785712588.0, "reward": 0.060546875, "reward_std": 0.11355571448802948, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 1857 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.677734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2023.0, "completions/mean_length": 1588.302734375, "completions/mean_terminated_length": 621.8484497070312, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6342920542801058, "frac_reward_zero_std": 0.71875, "grad_norm": 27.131050244956064, "learning_rate": 4.202539273212572e-07, "loss": 0.0066, "num_tokens": 1786602151.0, "reward": 0.0703125, "reward_std": 0.09529343992471695, "rewards/accuracy_reward/mean": 0.0703125, "rewards/accuracy_reward/std": 0.25592297315597534, "step": 1858 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1991.0, "completions/mean_length": 1636.376953125, "completions/mean_terminated_length": 661.4802856445312, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6346334385934966, "frac_reward_zero_std": 0.8125, "grad_norm": 16.303459305289422, "learning_rate": 4.197406786003569e-07, "loss": 0.0138, "num_tokens": 1787515336.0, "reward": 0.02734375, "reward_std": 0.06304662674665451, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 1859 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.630859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.0, "completions/mean_length": 1506.029296875, "completions/mean_terminated_length": 579.80419921875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.6349748229068874, "frac_reward_zero_std": 0.71875, "grad_norm": 24.651070077108315, "learning_rate": 4.192276147587387e-07, "loss": 0.0298, "num_tokens": 1788367799.0, "reward": 0.03515625, "reward_std": 0.0917677953839302, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 1860 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1971.0, "completions/mean_length": 1666.62890625, "completions/mean_terminated_length": 653.2714233398438, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.6353162072202783, "frac_reward_zero_std": 0.875, "grad_norm": 11.958992180749624, "learning_rate": 4.1871473652460265e-07, "loss": 0.0117, "num_tokens": 1789308969.0, "reward": 0.05078125, "reward_std": 0.054534729570150375, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 1861 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.62890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1924.0, "completions/mean_length": 1533.990234375, "completions/mean_terminated_length": 662.8789672851562, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6356575915336691, "frac_reward_zero_std": 0.78125, "grad_norm": 13.218555622655918, "learning_rate": 4.182020446258857e-07, "loss": 0.0193, "num_tokens": 1790169508.0, "reward": 0.060546875, "reward_std": 0.09742391854524612, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 1862 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.755859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1893.0, "completions/mean_length": 1701.494140625, "completions/mean_terminated_length": 628.7120361328125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.6359989758470598, "frac_reward_zero_std": 0.65625, "grad_norm": 17.92061487289862, "learning_rate": 4.1768953979026024e-07, "loss": 0.0278, "num_tokens": 1791117665.0, "reward": 0.0859375, "reward_std": 0.1312175691127777, "rewards/accuracy_reward/mean": 0.0859375, "rewards/accuracy_reward/std": 0.28054583072662354, "step": 1863 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.759765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1869.0, "completions/mean_length": 1714.470703125, "completions/mean_terminated_length": 659.650390625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6363403601604506, "frac_reward_zero_std": 0.75, "grad_norm": 22.221371870284877, "learning_rate": 4.171772227451331e-07, "loss": 0.0075, "num_tokens": 1792072786.0, "reward": 0.046875, "reward_std": 0.09413031488656998, "rewards/accuracy_reward/mean": 0.046875, "rewards/accuracy_reward/std": 0.21157780289649963, "step": 1864 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1957.0, "completions/mean_length": 1783.560546875, "completions/mean_terminated_length": 694.6099853515625, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.6366817444738414, "frac_reward_zero_std": 0.75, "grad_norm": 25.433143644358598, "learning_rate": 4.166650942176447e-07, "loss": 0.01, "num_tokens": 1793072257.0, "reward": 0.064453125, "reward_std": 0.09016359597444534, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 1865 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.76953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1817.0, "completions/mean_length": 1700.888671875, "completions/mean_terminated_length": 541.8898315429688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6370231287872322, "frac_reward_zero_std": 0.71875, "grad_norm": 85.78492015725827, "learning_rate": 4.1615315493466797e-07, "loss": 0.0126, "num_tokens": 1794017816.0, "reward": 0.03125, "reward_std": 0.08290977776050568, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 1866 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.74609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1971.0, "completions/mean_length": 1686.474609375, "completions/mean_terminated_length": 624.1461791992188, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.637364513100623, "frac_reward_zero_std": 0.84375, "grad_norm": 54.130651710185056, "learning_rate": 4.156414056228065e-07, "loss": 0.0062, "num_tokens": 1794957307.0, "reward": 0.041015625, "reward_std": 0.06288585811853409, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 1867 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.701171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2027.0, "completions/mean_length": 1615.66796875, "completions/mean_terminated_length": 601.2418212890625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6377058974140138, "frac_reward_zero_std": 0.59375, "grad_norm": 17.881941645674697, "learning_rate": 4.151298470083954e-07, "loss": 0.041, "num_tokens": 1795859041.0, "reward": 0.107421875, "reward_std": 0.13792334496974945, "rewards/accuracy_reward/mean": 0.107421875, "rewards/accuracy_reward/std": 0.30995169281959534, "step": 1868 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1983.0, "completions/mean_length": 1700.154296875, "completions/mean_terminated_length": 611.7338256835938, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6380472817274047, "frac_reward_zero_std": 0.75, "grad_norm": 24.588248207098324, "learning_rate": 4.146184798174983e-07, "loss": 0.0099, "num_tokens": 1796807584.0, "reward": 0.05078125, "reward_std": 0.09028453379869461, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 1869 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.759765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1939.0, "completions/mean_length": 1704.033203125, "completions/mean_terminated_length": 617.081298828125, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.6383886660407955, "frac_reward_zero_std": 0.875, "grad_norm": 45.12812514210717, "learning_rate": 4.141073047759076e-07, "loss": 0.011, "num_tokens": 1797771777.0, "reward": 0.025390625, "reward_std": 0.048406004905700684, "rewards/accuracy_reward/mean": 0.025390625, "rewards/accuracy_reward/std": 0.15746226906776428, "step": 1870 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.73046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1662.0, "completions/mean_length": 1635.787109375, "completions/mean_terminated_length": 518.6304321289062, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6387300503541862, "frac_reward_zero_std": 0.625, "grad_norm": 25.654035521992082, "learning_rate": 4.135963226091426e-07, "loss": 0.0101, "num_tokens": 1798690260.0, "reward": 0.107421875, "reward_std": 0.1643633246421814, "rewards/accuracy_reward/mean": 0.107421875, "rewards/accuracy_reward/std": 0.30995169281959534, "step": 1871 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1953.0, "completions/mean_length": 1610.197265625, "completions/mean_terminated_length": 647.03125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.639071434667577, "frac_reward_zero_std": 0.78125, "grad_norm": 20.45110421108804, "learning_rate": 4.1308553404244927e-07, "loss": 0.0148, "num_tokens": 1799589305.0, "reward": 0.064453125, "reward_std": 0.0829630047082901, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 1872 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1779.0, "completions/mean_length": 1676.05078125, "completions/mean_terminated_length": 605.2879028320312, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.6394128189809678, "frac_reward_zero_std": 0.71875, "grad_norm": 34.66645930680124, "learning_rate": 4.1257493980079825e-07, "loss": 0.0273, "num_tokens": 1800517619.0, "reward": 0.08984375, "reward_std": 0.10656822472810745, "rewards/accuracy_reward/mean": 0.08984375, "rewards/accuracy_reward/std": 0.2862374484539032, "step": 1873 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1847.0, "completions/mean_length": 1752.974609375, "completions/mean_terminated_length": 596.0769653320312, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.6397542032943586, "frac_reward_zero_std": 0.90625, "grad_norm": 2.2474519057455, "learning_rate": 4.120645406088846e-07, "loss": -0.0009, "num_tokens": 1801494662.0, "reward": 0.029296875, "reward_std": 0.028222277760505676, "rewards/accuracy_reward/mean": 0.030241934582591057, "rewards/accuracy_reward/std": 0.1714252382516861, "step": 1874 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.70703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1980.0, "completions/mean_length": 1606.20703125, "completions/mean_terminated_length": 540.0133666992188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6400955876077494, "frac_reward_zero_std": 0.78125, "grad_norm": 10.664033533125084, "learning_rate": 4.1155433719112696e-07, "loss": 0.0149, "num_tokens": 1802403056.0, "reward": 0.02734375, "reward_std": 0.06987475603818893, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 1875 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1944.0, "completions/mean_length": 1650.171875, "completions/mean_terminated_length": 593.085693359375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6404369719211402, "frac_reward_zero_std": 0.6875, "grad_norm": 17.07520178585204, "learning_rate": 4.1104433027166564e-07, "loss": 0.0374, "num_tokens": 1803318472.0, "reward": 0.056640625, "reward_std": 0.10837717354297638, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 1876 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.708984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1710.0, "completions/mean_length": 1637.01953125, "completions/mean_terminated_length": 635.7717895507812, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.640778356234531, "frac_reward_zero_std": 0.8125, "grad_norm": 53.899380724028475, "learning_rate": 4.1053452057436213e-07, "loss": 0.0118, "num_tokens": 1804233074.0, "reward": 0.033203125, "reward_std": 0.06920956075191498, "rewards/accuracy_reward/mean": 0.033203125, "rewards/accuracy_reward/std": 0.17934183776378632, "step": 1877 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.65234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1975.0, "completions/mean_length": 1555.181640625, "completions/mean_terminated_length": 630.455078125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6411197405479219, "frac_reward_zero_std": 0.84375, "grad_norm": 11.48270970018041, "learning_rate": 4.1002490882279804e-07, "loss": 0.0071, "num_tokens": 1805105887.0, "reward": 0.037109375, "reward_std": 0.064970001578331, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 1878 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.771484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2026.0, "completions/mean_length": 1716.28515625, "completions/mean_terminated_length": 596.3931884765625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.6414611248613126, "frac_reward_zero_std": 0.84375, "grad_norm": 18.68429825475646, "learning_rate": 4.0951549574027434e-07, "loss": 0.0088, "num_tokens": 1806052625.0, "reward": 0.025390625, "reward_std": 0.057656385004520416, "rewards/accuracy_reward/mean": 0.026209676638245583, "rewards/accuracy_reward/std": 0.1599196344614029, "step": 1879 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.75, "completions/max_length": 2048.0, "completions/max_terminated_length": 1701.0, "completions/mean_length": 1684.283203125, "completions/mean_terminated_length": 593.1328125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6418025091747034, "frac_reward_zero_std": 0.75, "grad_norm": 49.4379803828236, "learning_rate": 4.0900628204980924e-07, "loss": 0.0165, "num_tokens": 1807000402.0, "reward": 0.072265625, "reward_std": 0.08961933851242065, "rewards/accuracy_reward/mean": 0.072265625, "rewards/accuracy_reward/std": 0.2591804563999176, "step": 1880 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1928.0, "completions/mean_length": 1665.091796875, "completions/mean_terminated_length": 705.1986083984375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6421438934880942, "frac_reward_zero_std": 0.75, "grad_norm": 27.902747443619923, "learning_rate": 4.084972684741386e-07, "loss": 0.0192, "num_tokens": 1807929857.0, "reward": 0.0546875, "reward_std": 0.09880206733942032, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 1881 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.66796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1899.0, "completions/mean_length": 1559.21484375, "completions/mean_terminated_length": 575.8941040039062, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.642485277801485, "frac_reward_zero_std": 0.6875, "grad_norm": 51.563566395950225, "learning_rate": 4.079884557357142e-07, "loss": 0.009, "num_tokens": 1808809215.0, "reward": 0.041015625, "reward_std": 0.10100477933883667, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 1882 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.68359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.0, "completions/mean_length": 1618.8984375, "completions/mean_terminated_length": 691.8271484375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6428266621148758, "frac_reward_zero_std": 0.71875, "grad_norm": 18.205882448412563, "learning_rate": 4.0747984455670257e-07, "loss": 0.0082, "num_tokens": 1809719483.0, "reward": 0.06640625, "reward_std": 0.11052703112363815, "rewards/accuracy_reward/mean": 0.06640625, "rewards/accuracy_reward/std": 0.2492343932390213, "step": 1883 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1874.0, "completions/mean_length": 1590.919921875, "completions/mean_terminated_length": 547.8397216796875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6431680464282666, "frac_reward_zero_std": 0.78125, "grad_norm": 18.997819879991024, "learning_rate": 4.069714356589844e-07, "loss": 0.0026, "num_tokens": 1810611906.0, "reward": 0.064453125, "reward_std": 0.08644992858171463, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 1884 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1986.0, "completions/mean_length": 1558.3671875, "completions/mean_terminated_length": 655.2666625976562, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6435094307416575, "frac_reward_zero_std": 0.84375, "grad_norm": 5.680236661237201, "learning_rate": 4.064632297641533e-07, "loss": 0.0091, "num_tokens": 1811491502.0, "reward": 0.037109375, "reward_std": 0.0557793527841568, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 1885 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2028.0, "completions/mean_length": 1693.7265625, "completions/mean_terminated_length": 714.2647094726562, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.6438508150550483, "frac_reward_zero_std": 0.65625, "grad_norm": 23.44396970075306, "learning_rate": 4.05955227593514e-07, "loss": 0.0029, "num_tokens": 1812434962.0, "reward": 0.0859375, "reward_std": 0.1332087218761444, "rewards/accuracy_reward/mean": 0.0859375, "rewards/accuracy_reward/std": 0.28054583072662354, "step": 1886 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2028.0, "completions/mean_length": 1624.328125, "completions/mean_terminated_length": 620.8947143554688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.644192199368439, "frac_reward_zero_std": 0.75, "grad_norm": 21.949686178829488, "learning_rate": 4.05447429868083e-07, "loss": 0.0092, "num_tokens": 1813350810.0, "reward": 0.06640625, "reward_std": 0.09247933328151703, "rewards/accuracy_reward/mean": 0.06640625, "rewards/accuracy_reward/std": 0.2492343932390213, "step": 1887 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1992.0, "completions/mean_length": 1635.7421875, "completions/mean_terminated_length": 659.3421020507812, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.6445335836818298, "frac_reward_zero_std": 0.875, "grad_norm": 64.40457654317638, "learning_rate": 4.049398373085862e-07, "loss": 0.01, "num_tokens": 1814262134.0, "reward": 0.033203125, "reward_std": 0.04219770431518555, "rewards/accuracy_reward/mean": 0.033203125, "rewards/accuracy_reward/std": 0.17934183776378632, "step": 1888 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.67578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1899.0, "completions/mean_length": 1607.421875, "completions/mean_terminated_length": 689.1083984375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6448749679952206, "frac_reward_zero_std": 0.84375, "grad_norm": 33.52260018606143, "learning_rate": 4.044324506354585e-07, "loss": 0.0172, "num_tokens": 1815162846.0, "reward": 0.03125, "reward_std": 0.06068410724401474, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 1889 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.62109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2013.0, "completions/mean_length": 1494.90234375, "completions/mean_terminated_length": 588.2783203125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.6452163523086114, "frac_reward_zero_std": 0.6875, "grad_norm": 14.791239591765395, "learning_rate": 4.0392527056884254e-07, "loss": 0.0232, "num_tokens": 1815999036.0, "reward": 0.044921875, "reward_std": 0.10453138500452042, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 1890 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1806.0, "completions/mean_length": 1620.541015625, "completions/mean_terminated_length": 645.0576782226562, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6455577366220022, "frac_reward_zero_std": 0.6875, "grad_norm": 39.13802754860844, "learning_rate": 4.034182978285877e-07, "loss": 0.0249, "num_tokens": 1816906577.0, "reward": 0.087890625, "reward_std": 0.13353893160820007, "rewards/accuracy_reward/mean": 0.087890625, "rewards/accuracy_reward/std": 0.2834126651287079, "step": 1891 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.658203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1926.0, "completions/mean_length": 1547.8671875, "completions/mean_terminated_length": 585.0171508789062, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.645899120935393, "frac_reward_zero_std": 0.875, "grad_norm": 7.715313124520083, "learning_rate": 4.0291153313424874e-07, "loss": -0.005, "num_tokens": 1817777373.0, "reward": 0.025390625, "reward_std": 0.047094546258449554, "rewards/accuracy_reward/mean": 0.026209676638245583, "rewards/accuracy_reward/std": 0.1599196344614029, "step": 1892 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.69921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1914.0, "completions/mean_length": 1626.55078125, "completions/mean_terminated_length": 646.8181762695312, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6462405052487838, "frac_reward_zero_std": 0.65625, "grad_norm": 21.383252866922586, "learning_rate": 4.024049772050857e-07, "loss": 0.0105, "num_tokens": 1818693415.0, "reward": 0.052734375, "reward_std": 0.10854348540306091, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 1893 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.69140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2008.0, "completions/mean_length": 1603.267578125, "completions/mean_terminated_length": 606.841796875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6465818895621747, "frac_reward_zero_std": 0.84375, "grad_norm": 12.602934390656614, "learning_rate": 4.01898630760062e-07, "loss": 0.0218, "num_tokens": 1819587568.0, "reward": 0.052734375, "reward_std": 0.06260652095079422, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 1894 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.716796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2018.0, "completions/mean_length": 1662.603515625, "completions/mean_terminated_length": 687.1517333984375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6469232738755654, "frac_reward_zero_std": 0.6875, "grad_norm": 75.83647106595872, "learning_rate": 4.0139249451784383e-07, "loss": 0.0135, "num_tokens": 1820519669.0, "reward": 0.0625, "reward_std": 0.11294380575418472, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.2422981858253479, "step": 1895 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.759765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2048.0, "completions/mean_length": 1742.587890625, "completions/mean_terminated_length": 777.6422729492188, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.6472646581889562, "frac_reward_zero_std": 0.65625, "grad_norm": 38.37533702448189, "learning_rate": 4.00886569196799e-07, "loss": 0.0223, "num_tokens": 1821497842.0, "reward": 0.0703125, "reward_std": 0.13247431814670563, "rewards/accuracy_reward/mean": 0.0703125, "rewards/accuracy_reward/std": 0.25592297315597534, "step": 1896 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.654296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2002.0, "completions/mean_length": 1578.185546875, "completions/mean_terminated_length": 688.9887084960938, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.647606042502347, "frac_reward_zero_std": 0.6875, "grad_norm": 28.96005356889781, "learning_rate": 4.003808555149961e-07, "loss": 0.0146, "num_tokens": 1822383697.0, "reward": 0.08203125, "reward_std": 0.12285147607326508, "rewards/accuracy_reward/mean": 0.08203125, "rewards/accuracy_reward/std": 0.2746807038784027, "step": 1897 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1965.0, "completions/mean_length": 1606.001953125, "completions/mean_terminated_length": 700.9583740234375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.6479474268157378, "frac_reward_zero_std": 0.875, "grad_norm": 18.487891776182312, "learning_rate": 3.9987535419020303e-07, "loss": 0.0269, "num_tokens": 1823281090.0, "reward": 0.0546875, "reward_std": 0.059858135879039764, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 1898 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.69921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1956.0, "completions/mean_length": 1641.009765625, "completions/mean_terminated_length": 694.8895874023438, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.6482888111291286, "frac_reward_zero_std": 0.84375, "grad_norm": 70.17150869471854, "learning_rate": 3.993700659398863e-07, "loss": -0.0089, "num_tokens": 1824197991.0, "reward": 0.046875, "reward_std": 0.057157501578330994, "rewards/accuracy_reward/mean": 0.046875, "rewards/accuracy_reward/std": 0.21157780289649963, "step": 1899 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.677734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2007.0, "completions/mean_length": 1602.35546875, "completions/mean_terminated_length": 665.1514892578125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6486301954425194, "frac_reward_zero_std": 0.625, "grad_norm": 56.92204271812513, "learning_rate": 3.9886499148121055e-07, "loss": 0.0284, "num_tokens": 1825098925.0, "reward": 0.107421875, "reward_std": 0.1464010328054428, "rewards/accuracy_reward/mean": 0.107421875, "rewards/accuracy_reward/std": 0.30995169281959534, "step": 1900 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1967.0, "completions/mean_length": 1608.669921875, "completions/mean_terminated_length": 568.1513061523438, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6489715797559102, "frac_reward_zero_std": 0.78125, "grad_norm": 28.794096151075603, "learning_rate": 3.9836013153103643e-07, "loss": 0.029, "num_tokens": 1825998420.0, "reward": 0.0546875, "reward_std": 0.07861737906932831, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 1901 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.658203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2005.0, "completions/mean_length": 1571.388671875, "completions/mean_terminated_length": 653.8856811523438, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6493129640693011, "frac_reward_zero_std": 0.65625, "grad_norm": 40.29705204654067, "learning_rate": 3.9785548680592027e-07, "loss": 0.0086, "num_tokens": 1826879147.0, "reward": 0.08984375, "reward_std": 0.1296786367893219, "rewards/accuracy_reward/mean": 0.08984375, "rewards/accuracy_reward/std": 0.2862374484539032, "step": 1902 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.685546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2034.0, "completions/mean_length": 1604.296875, "completions/mean_terminated_length": 636.9689331054688, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.6496543483826919, "frac_reward_zero_std": 0.71875, "grad_norm": 18.014936877204462, "learning_rate": 3.97351058022113e-07, "loss": 0.0019, "num_tokens": 1827780131.0, "reward": 0.0625, "reward_std": 0.09435540437698364, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.2422981858253479, "step": 1903 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.72265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1881.0, "completions/mean_length": 1678.91796875, "completions/mean_terminated_length": 717.225341796875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.6499957326960826, "frac_reward_zero_std": 0.75, "grad_norm": 21.39527370249335, "learning_rate": 3.9684684589555894e-07, "loss": 0.0282, "num_tokens": 1828719065.0, "reward": 0.03515625, "reward_std": 0.08345641195774078, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 1904 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.673828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2040.0, "completions/mean_length": 1605.140625, "completions/mean_terminated_length": 690.2515258789062, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6503371170094734, "frac_reward_zero_std": 0.8125, "grad_norm": 22.325634809261118, "learning_rate": 3.9634285114189505e-07, "loss": -0.0177, "num_tokens": 1829619265.0, "reward": 0.041015625, "reward_std": 0.062062256038188934, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 1905 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2019.0, "completions/mean_length": 1625.32421875, "completions/mean_terminated_length": 660.7564086914062, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6506785013228642, "frac_reward_zero_std": 0.6875, "grad_norm": 61.240552491319065, "learning_rate": 3.958390744764497e-07, "loss": 0.0201, "num_tokens": 1830529255.0, "reward": 0.05859375, "reward_std": 0.10150270164012909, "rewards/accuracy_reward/mean": 0.05859375, "rewards/accuracy_reward/std": 0.23509246110916138, "step": 1906 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.72265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1874.0, "completions/mean_length": 1626.369140625, "completions/mean_terminated_length": 527.7535400390625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.651019885636255, "frac_reward_zero_std": 0.78125, "grad_norm": 10.628270906804987, "learning_rate": 3.953355166142417e-07, "loss": -0.0007, "num_tokens": 1831440260.0, "reward": 0.01953125, "reward_std": 0.06233368441462517, "rewards/accuracy_reward/mean": 0.02083333395421505, "rewards/accuracy_reward/std": 0.14297515153884888, "step": 1907 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.66015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2023.0, "completions/mean_length": 1572.7890625, "completions/mean_terminated_length": 649.6781616210938, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6513612699496458, "frac_reward_zero_std": 0.65625, "grad_norm": 15.521275011062448, "learning_rate": 3.9483217826997927e-07, "loss": 0.0316, "num_tokens": 1832317544.0, "reward": 0.0859375, "reward_std": 0.12324429303407669, "rewards/accuracy_reward/mean": 0.0859375, "rewards/accuracy_reward/std": 0.28054583072662354, "step": 1908 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.689453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2029.0, "completions/mean_length": 1608.9921875, "completions/mean_terminated_length": 634.5848999023438, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.6517026542630366, "frac_reward_zero_std": 0.8125, "grad_norm": 5.360901834856371, "learning_rate": 3.9432906015805946e-07, "loss": 0.01, "num_tokens": 1833225940.0, "reward": 0.0234375, "reward_std": 0.062167368829250336, "rewards/accuracy_reward/mean": 0.0234375, "rewards/accuracy_reward/std": 0.15143637359142303, "step": 1909 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.611328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1926.0, "completions/mean_length": 1465.458984375, "completions/mean_terminated_length": 549.2009887695312, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.6520440385764275, "frac_reward_zero_std": 0.84375, "grad_norm": 27.16075614590475, "learning_rate": 3.93826162992566e-07, "loss": -0.0027, "num_tokens": 1834055327.0, "reward": 0.046875, "reward_std": 0.059858135879039764, "rewards/accuracy_reward/mean": 0.046875, "rewards/accuracy_reward/std": 0.21157780289649963, "step": 1910 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.646484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2031.0, "completions/mean_length": 1554.236328125, "completions/mean_terminated_length": 651.2762451171875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.6523854228898183, "frac_reward_zero_std": 0.78125, "grad_norm": 11.83903793745618, "learning_rate": 3.933234874872695e-07, "loss": 0.0198, "num_tokens": 1834927048.0, "reward": 0.03515625, "reward_std": 0.07796105742454529, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 1911 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.576171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1501.0, "completions/mean_length": 1427.958984375, "completions/mean_terminated_length": 585.0460815429688, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.652726807203209, "frac_reward_zero_std": 0.71875, "grad_norm": 2.8231449167928226, "learning_rate": 3.92821034355626e-07, "loss": 0.0282, "num_tokens": 1835731171.0, "reward": 0.076171875, "reward_std": 0.11377984285354614, "rewards/accuracy_reward/mean": 0.076171875, "rewards/accuracy_reward/std": 0.26553234457969666, "step": 1912 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.62890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2003.0, "completions/mean_length": 1531.041015625, "completions/mean_terminated_length": 654.9315795898438, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.6530681915165998, "frac_reward_zero_std": 0.71875, "grad_norm": 12.60456839287929, "learning_rate": 3.923188043107758e-07, "loss": 0.0473, "num_tokens": 1836590040.0, "reward": 0.048828125, "reward_std": 0.09601948410272598, "rewards/accuracy_reward/mean": 0.05040322616696358, "rewards/accuracy_reward/std": 0.21899642050266266, "step": 1913 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.708984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1959.0, "completions/mean_length": 1647.740234375, "completions/mean_terminated_length": 672.6107177734375, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.6534095758299906, "frac_reward_zero_std": 0.78125, "grad_norm": 41.09799395784116, "learning_rate": 3.9181679806554267e-07, "loss": 0.0209, "num_tokens": 1837509219.0, "reward": 0.080078125, "reward_std": 0.09534214437007904, "rewards/accuracy_reward/mean": 0.08266129344701767, "rewards/accuracy_reward/std": 0.2756475806236267, "step": 1914 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.69921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1828.0, "completions/mean_length": 1625.287109375, "completions/mean_terminated_length": 642.6168823242188, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.6537509601433814, "frac_reward_zero_std": 0.75, "grad_norm": 37.618429167559775, "learning_rate": 3.913150163324326e-07, "loss": 0.017, "num_tokens": 1838416566.0, "reward": 0.0546875, "reward_std": 0.0767945945262909, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 1915 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1950.0, "completions/mean_length": 1529.771484375, "completions/mean_terminated_length": 573.9277954101562, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.6540923444567722, "frac_reward_zero_std": 0.875, "grad_norm": 0.21945231270445678, "learning_rate": 3.90813459823633e-07, "loss": 0.0099, "num_tokens": 1839278673.0, "reward": 0.06640625, "reward_std": 0.04847269132733345, "rewards/accuracy_reward/mean": 0.06640625, "rewards/accuracy_reward/std": 0.2492343932390213, "step": 1916 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.623046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1890.0, "completions/mean_length": 1557.873046875, "completions/mean_terminated_length": 747.766845703125, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.654433728770163, "frac_reward_zero_std": 0.71875, "grad_norm": 1.610379892229922, "learning_rate": 3.9031212925101144e-07, "loss": 0.0204, "num_tokens": 1840152768.0, "reward": 0.04296875, "reward_std": 0.09671889245510101, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 1917 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.677734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1949.0, "completions/mean_length": 1597.68359375, "completions/mean_terminated_length": 650.654541015625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.6547751130835539, "frac_reward_zero_std": 0.75, "grad_norm": 27.946039753780898, "learning_rate": 3.898110253261151e-07, "loss": 0.0129, "num_tokens": 1841048654.0, "reward": 0.06640625, "reward_std": 0.09580586850643158, "rewards/accuracy_reward/mean": 0.06854838877916336, "rewards/accuracy_reward/std": 0.25293970108032227, "step": 1918 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.673828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1863.0, "completions/mean_length": 1577.87109375, "completions/mean_terminated_length": 606.646728515625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.6551164973969447, "frac_reward_zero_std": 0.625, "grad_norm": 10.168538975497906, "learning_rate": 3.8931014876016944e-07, "loss": 0.0135, "num_tokens": 1841936076.0, "reward": 0.0703125, "reward_std": 0.12538030743598938, "rewards/accuracy_reward/mean": 0.0703125, "rewards/accuracy_reward/std": 0.25592297315597534, "step": 1919 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.63671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1996.0, "completions/mean_length": 1566.767578125, "completions/mean_terminated_length": 723.3171997070312, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.6554578817103354, "frac_reward_zero_std": 0.71875, "grad_norm": 12.76728289799827, "learning_rate": 3.88809500264077e-07, "loss": 0.0194, "num_tokens": 1842813733.0, "reward": 0.060546875, "reward_std": 0.10441835969686508, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.24230584502220154, "step": 1920 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.650390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1977.0, "completions/mean_length": 1570.7421875, "completions/mean_terminated_length": 682.8826904296875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6557992660237262, "frac_reward_zero_std": 0.8125, "grad_norm": 12.947854748541362, "learning_rate": 3.8830908054841673e-07, "loss": -0.0077, "num_tokens": 1843694481.0, "reward": 0.02734375, "reward_std": 0.05806133896112442, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 1921 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.724609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1707.0, "completions/mean_length": 1662.904296875, "completions/mean_terminated_length": 649.6383056640625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.656140650337117, "frac_reward_zero_std": 0.75, "grad_norm": 8.282081646655259, "learning_rate": 3.878088903234431e-07, "loss": 0.0317, "num_tokens": 1844626656.0, "reward": 0.04296875, "reward_std": 0.09248073399066925, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 1922 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2012.0, "completions/mean_length": 1629.91015625, "completions/mean_terminated_length": 803.4534912109375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6564820346505078, "frac_reward_zero_std": 0.875, "grad_norm": 20.364429226048358, "learning_rate": 3.873089302990844e-07, "loss": 0.005, "num_tokens": 1845553026.0, "reward": 0.025390625, "reward_std": 0.04957009106874466, "rewards/accuracy_reward/mean": 0.025390625, "rewards/accuracy_reward/std": 0.15746226906776428, "step": 1923 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.67578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1887.0, "completions/mean_length": 1578.521484375, "completions/mean_terminated_length": 599.9698486328125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6568234189638986, "frac_reward_zero_std": 0.78125, "grad_norm": 114.6146550085507, "learning_rate": 3.868092011849425e-07, "loss": 0.0113, "num_tokens": 1846431917.0, "reward": 0.041015625, "reward_std": 0.0814797505736351, "rewards/accuracy_reward/mean": 0.04233871027827263, "rewards/accuracy_reward/std": 0.2015640139579773, "step": 1924 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.626953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2042.0, "completions/mean_length": 1538.7109375, "completions/mean_terminated_length": 682.7853393554688, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6571648032772894, "frac_reward_zero_std": 0.71875, "grad_norm": 44.44046880427549, "learning_rate": 3.8630970369029146e-07, "loss": 0.0267, "num_tokens": 1847295129.0, "reward": 0.0625, "reward_std": 0.09959384053945541, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.2422981858253479, "step": 1925 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1927.0, "completions/mean_length": 1677.900390625, "completions/mean_terminated_length": 654.683837890625, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.6575061875906802, "frac_reward_zero_std": 0.71875, "grad_norm": 129.48153553409983, "learning_rate": 3.858104385240768e-07, "loss": 0.0164, "num_tokens": 1848230118.0, "reward": 0.029296875, "reward_std": 0.08654290437698364, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 1926 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.66796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2037.0, "completions/mean_length": 1573.306640625, "completions/mean_terminated_length": 618.3353271484375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6578475719040711, "frac_reward_zero_std": 0.78125, "grad_norm": 16.877465650830253, "learning_rate": 3.85311406394914e-07, "loss": 0.0014, "num_tokens": 1849109955.0, "reward": 0.078125, "reward_std": 0.0921671986579895, "rewards/accuracy_reward/mean": 0.078125, "rewards/accuracy_reward/std": 0.26863065361976624, "step": 1927 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.67578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2008.0, "completions/mean_length": 1601.75390625, "completions/mean_terminated_length": 671.62646484375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6581889562174618, "frac_reward_zero_std": 0.71875, "grad_norm": 26.7497106550931, "learning_rate": 3.848126080110878e-07, "loss": 0.0149, "num_tokens": 1850018917.0, "reward": 0.0546875, "reward_std": 0.10381289571523666, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 1928 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.72265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1948.0, "completions/mean_length": 1643.60546875, "completions/mean_terminated_length": 589.9013671875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.6585303405308526, "frac_reward_zero_std": 0.75, "grad_norm": 17.43784689102482, "learning_rate": 3.8431404408055133e-07, "loss": 0.0153, "num_tokens": 1850943947.0, "reward": 0.048828125, "reward_std": 0.08395528793334961, "rewards/accuracy_reward/mean": 0.05040322616696358, "rewards/accuracy_reward/std": 0.21899640560150146, "step": 1929 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.759765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1841.0, "completions/mean_length": 1705.958984375, "completions/mean_terminated_length": 624.219482421875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6588717248442434, "frac_reward_zero_std": 0.8125, "grad_norm": 13.649750323093194, "learning_rate": 3.8381571531092496e-07, "loss": 0.0111, "num_tokens": 1851896070.0, "reward": 0.029296875, "reward_std": 0.059984706342220306, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 1930 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.681640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1974.0, "completions/mean_length": 1604.974609375, "completions/mean_terminated_length": 656.4110107421875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.6592131091576342, "frac_reward_zero_std": 0.84375, "grad_norm": 12.276748754782338, "learning_rate": 3.8331762240949503e-07, "loss": 0.0119, "num_tokens": 1852797305.0, "reward": 0.01953125, "reward_std": 0.04907120764255524, "rewards/accuracy_reward/mean": 0.01953125, "rewards/accuracy_reward/std": 0.1385180652141571, "step": 1931 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1817.0, "completions/mean_length": 1666.044921875, "completions/mean_terminated_length": 689.9375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.659554493471025, "frac_reward_zero_std": 0.71875, "grad_norm": 38.474988449109986, "learning_rate": 3.8281976608321366e-07, "loss": 0.0273, "num_tokens": 1853723456.0, "reward": 0.07421875, "reward_std": 0.11305920034646988, "rewards/accuracy_reward/mean": 0.07421875, "rewards/accuracy_reward/std": 0.2623828947544098, "step": 1932 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.70703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2018.0, "completions/mean_length": 1644.75, "completions/mean_terminated_length": 671.5733642578125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6598958777844158, "frac_reward_zero_std": 0.84375, "grad_norm": 29.958146312958622, "learning_rate": 3.823221470386965e-07, "loss": 0.0123, "num_tokens": 1854641200.0, "reward": 0.021484375, "reward_std": 0.05193261057138443, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 1933 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.693359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1720.0, "completions/mean_length": 1613.302734375, "completions/mean_terminated_length": 630.3885498046875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6602372620978066, "frac_reward_zero_std": 0.71875, "grad_norm": 134.38455745394924, "learning_rate": 3.81824765982223e-07, "loss": 0.0131, "num_tokens": 1855542315.0, "reward": 0.0625, "reward_std": 0.11360110342502594, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.2422981858253479, "step": 1934 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.77734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1831.0, "completions/mean_length": 1734.376953125, "completions/mean_terminated_length": 639.4473876953125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6605786464111975, "frac_reward_zero_std": 0.8125, "grad_norm": 12.146875587259746, "learning_rate": 3.8132762361973456e-07, "loss": -0.001, "num_tokens": 1856506956.0, "reward": 0.021484375, "reward_std": 0.0546875, "rewards/accuracy_reward/mean": 0.02217741869390011, "rewards/accuracy_reward/std": 0.14740893244743347, "step": 1935 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.69140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1964.0, "completions/mean_length": 1624.94140625, "completions/mean_terminated_length": 677.0759887695312, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6609200307245882, "frac_reward_zero_std": 0.65625, "grad_norm": 20.820176880599945, "learning_rate": 3.8083072065683373e-07, "loss": 0.031, "num_tokens": 1857415870.0, "reward": 0.068359375, "reward_std": 0.12542080879211426, "rewards/accuracy_reward/mean": 0.068359375, "rewards/accuracy_reward/std": 0.25260838866233826, "step": 1936 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1702.0, "completions/mean_length": 1646.38671875, "completions/mean_terminated_length": 639.6027221679688, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.661261415037979, "frac_reward_zero_std": 0.71875, "grad_norm": 1.842423646244592, "learning_rate": 3.803340577987834e-07, "loss": 0.0034, "num_tokens": 1858335076.0, "reward": 0.041015625, "reward_std": 0.09291848540306091, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 1937 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.70703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2023.0, "completions/mean_length": 1596.2734375, "completions/mean_terminated_length": 506.106689453125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6616027993513698, "frac_reward_zero_std": 0.625, "grad_norm": 129.0727114858703, "learning_rate": 3.7983763575050575e-07, "loss": -0.0052, "num_tokens": 1859236000.0, "reward": 0.06640625, "reward_std": 0.11382715404033661, "rewards/accuracy_reward/mean": 0.06854838877916336, "rewards/accuracy_reward/std": 0.25293973088264465, "step": 1938 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2022.0, "completions/mean_length": 1600.294921875, "completions/mean_terminated_length": 650.2865600585938, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6619441836647606, "frac_reward_zero_std": 0.75, "grad_norm": 15.79687851534851, "learning_rate": 3.79341455216581e-07, "loss": 0.0119, "num_tokens": 1860129351.0, "reward": 0.04296875, "reward_std": 0.0801548883318901, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 1939 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.69921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1844.0, "completions/mean_length": 1594.65625, "completions/mean_terminated_length": 540.7792358398438, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6622855679781514, "frac_reward_zero_std": 0.8125, "grad_norm": 21.302880074106984, "learning_rate": 3.788455169012469e-07, "loss": 0.008, "num_tokens": 1861015591.0, "reward": 0.03515625, "reward_std": 0.06888246536254883, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 1940 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.67578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2029.0, "completions/mean_length": 1591.091796875, "completions/mean_terminated_length": 638.740966796875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6626269522915422, "frac_reward_zero_std": 0.65625, "grad_norm": 27.62542811601489, "learning_rate": 3.783498215083967e-07, "loss": 0.0229, "num_tokens": 1861903622.0, "reward": 0.0625, "reward_std": 0.11179865896701813, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.2422981858253479, "step": 1941 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1996.0, "completions/mean_length": 1573.845703125, "completions/mean_terminated_length": 728.6141357421875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.662968336604933, "frac_reward_zero_std": 0.59375, "grad_norm": 113.42485994310648, "learning_rate": 3.778543697415797e-07, "loss": 0.0354, "num_tokens": 1862778471.0, "reward": 0.10546875, "reward_std": 0.13864420354366302, "rewards/accuracy_reward/mean": 0.10546875, "rewards/accuracy_reward/std": 0.3074568510055542, "step": 1942 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.708984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.0, "completions/mean_length": 1670.068359375, "completions/mean_terminated_length": 749.671142578125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6633097209183239, "frac_reward_zero_std": 0.78125, "grad_norm": 25.13782941878759, "learning_rate": 3.77359162303999e-07, "loss": 0.0042, "num_tokens": 1863716378.0, "reward": 0.068359375, "reward_std": 0.09731185436248779, "rewards/accuracy_reward/mean": 0.068359375, "rewards/accuracy_reward/std": 0.25260838866233826, "step": 1943 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.68359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1824.0, "completions/mean_length": 1615.54296875, "completions/mean_terminated_length": 681.2222290039062, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6636511052317146, "frac_reward_zero_std": 0.6875, "grad_norm": 42.91499210315436, "learning_rate": 3.7686419989851104e-07, "loss": 0.0292, "num_tokens": 1864613968.0, "reward": 0.095703125, "reward_std": 0.13276061415672302, "rewards/accuracy_reward/mean": 0.095703125, "rewards/accuracy_reward/std": 0.2944713830947876, "step": 1944 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.716796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2043.0, "completions/mean_length": 1662.95703125, "completions/mean_terminated_length": 688.4000244140625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.6639924895451054, "frac_reward_zero_std": 0.84375, "grad_norm": 16.606609122648123, "learning_rate": 3.763694832276242e-07, "loss": -0.0075, "num_tokens": 1865541178.0, "reward": 0.03515625, "reward_std": 0.05644455552101135, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 1945 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.720703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1919.0, "completions/mean_length": 1673.373046875, "completions/mean_terminated_length": 706.6782836914062, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6643338738584962, "frac_reward_zero_std": 0.8125, "grad_norm": 10.341940137915266, "learning_rate": 3.758750129934988e-07, "loss": 0.0226, "num_tokens": 1866479097.0, "reward": 0.0234375, "reward_std": 0.05880707502365112, "rewards/accuracy_reward/mean": 0.0234375, "rewards/accuracy_reward/std": 0.15143637359142303, "step": 1946 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1990.0, "completions/mean_length": 1722.248046875, "completions/mean_terminated_length": 610.6034545898438, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.664675258171887, "frac_reward_zero_std": 0.8125, "grad_norm": 75.71014409972499, "learning_rate": 3.753807898979442e-07, "loss": -0.0091, "num_tokens": 1867441000.0, "reward": 0.03125, "reward_std": 0.06711846590042114, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 1947 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.67578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1999.0, "completions/mean_length": 1612.03515625, "completions/mean_terminated_length": 703.3373413085938, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6650166424852778, "frac_reward_zero_std": 0.59375, "grad_norm": 91.87204973098542, "learning_rate": 3.748868146424201e-07, "loss": -0.0058, "num_tokens": 1868343882.0, "reward": 0.060546875, "reward_std": 0.1333048790693283, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 1948 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.73046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1948.0, "completions/mean_length": 1678.384765625, "completions/mean_terminated_length": 676.6739501953125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.6653580267986686, "frac_reward_zero_std": 0.65625, "grad_norm": 46.45154082282707, "learning_rate": 3.7439308792803405e-07, "loss": 0.0092, "num_tokens": 1869281375.0, "reward": 0.099609375, "reward_std": 0.13248127698898315, "rewards/accuracy_reward/mean": 0.099609375, "rewards/accuracy_reward/std": 0.29977133870124817, "step": 1949 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.595703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2030.0, "completions/mean_length": 1494.88671875, "completions/mean_terminated_length": 679.9130249023438, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6656994111120594, "frac_reward_zero_std": 0.71875, "grad_norm": 48.06504360074654, "learning_rate": 3.738996104555406e-07, "loss": 0.0355, "num_tokens": 1870129173.0, "reward": 0.03515625, "reward_std": 0.08615703880786896, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 1950 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1940.0, "completions/mean_length": 1682.36328125, "completions/mean_terminated_length": 630.23486328125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6660407954254502, "frac_reward_zero_std": 0.78125, "grad_norm": 13.669517583765767, "learning_rate": 3.734063829253411e-07, "loss": 0.0034, "num_tokens": 1871066831.0, "reward": 0.029296875, "reward_std": 0.07328139245510101, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 1951 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.73828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2010.0, "completions/mean_length": 1677.826171875, "completions/mean_terminated_length": 633.962646484375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.666382179738841, "frac_reward_zero_std": 0.625, "grad_norm": 58.895488730141004, "learning_rate": 3.7291340603748146e-07, "loss": 0.0088, "num_tokens": 1872003782.0, "reward": 0.08984375, "reward_std": 0.12793609499931335, "rewards/accuracy_reward/mean": 0.08984375, "rewards/accuracy_reward/std": 0.2862374484539032, "step": 1952 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.767578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1963.0, "completions/mean_length": 1717.095703125, "completions/mean_terminated_length": 624.27734375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6667235640522318, "frac_reward_zero_std": 0.78125, "grad_norm": 56.18263103415097, "learning_rate": 3.724206804916526e-07, "loss": 0.0106, "num_tokens": 1872961031.0, "reward": 0.06640625, "reward_std": 0.08627147972583771, "rewards/accuracy_reward/mean": 0.06640625, "rewards/accuracy_reward/std": 0.2492343932390213, "step": 1953 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.685546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1940.0, "completions/mean_length": 1616.974609375, "completions/mean_terminated_length": 677.2857055664062, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.6670649483656226, "frac_reward_zero_std": 0.8125, "grad_norm": 16.378567912438644, "learning_rate": 3.7192820698718797e-07, "loss": 0.0085, "num_tokens": 1873875210.0, "reward": 0.021484375, "reward_std": 0.058760736137628555, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 1954 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.646484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1856.0, "completions/mean_length": 1535.353515625, "completions/mean_terminated_length": 597.8618774414062, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.6674063326790134, "frac_reward_zero_std": 0.8125, "grad_norm": 25.0177318452696, "learning_rate": 3.7143598622306374e-07, "loss": 0.0066, "num_tokens": 1874739039.0, "reward": 0.044921875, "reward_std": 0.07803337275981903, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 1955 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.681640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1868.0, "completions/mean_length": 1590.146484375, "completions/mean_terminated_length": 609.8343505859375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6677477169924042, "frac_reward_zero_std": 0.78125, "grad_norm": 40.91140456636508, "learning_rate": 3.7094401889789715e-07, "loss": 0.0101, "num_tokens": 1875630522.0, "reward": 0.07421875, "reward_std": 0.09367191791534424, "rewards/accuracy_reward/mean": 0.07421875, "rewards/accuracy_reward/std": 0.2623828947544098, "step": 1956 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1931.0, "completions/mean_length": 1636.171875, "completions/mean_terminated_length": 696.3590087890625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.668089101305795, "frac_reward_zero_std": 0.78125, "grad_norm": 52.58977840071835, "learning_rate": 3.70452305709946e-07, "loss": 0.0128, "num_tokens": 1876545618.0, "reward": 0.037109375, "reward_std": 0.07581022381782532, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 1957 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.716796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1944.0, "completions/mean_length": 1634.177734375, "completions/mean_terminated_length": 586.779296875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6684304856191858, "frac_reward_zero_std": 0.75, "grad_norm": 27.987811161614058, "learning_rate": 3.699608473571072e-07, "loss": -0.0051, "num_tokens": 1877455101.0, "reward": 0.0859375, "reward_std": 0.09811057895421982, "rewards/accuracy_reward/mean": 0.0859375, "rewards/accuracy_reward/std": 0.28054583072662354, "step": 1958 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2011.0, "completions/mean_length": 1639.21484375, "completions/mean_terminated_length": 671.0394897460938, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6687718699325766, "frac_reward_zero_std": 0.8125, "grad_norm": 15.322672873283942, "learning_rate": 3.694696445369161e-07, "loss": 0.0082, "num_tokens": 1878367179.0, "reward": 0.029296875, "reward_std": 0.06271953880786896, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 1959 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.771484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1996.0, "completions/mean_length": 1724.814453125, "completions/mean_terminated_length": 633.7179565429688, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6691132542459673, "frac_reward_zero_std": 0.6875, "grad_norm": 52.968759319113815, "learning_rate": 3.68978697946545e-07, "loss": -0.0021, "num_tokens": 1879338364.0, "reward": 0.0390625, "reward_std": 0.10056467354297638, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 1960 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.66015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2020.0, "completions/mean_length": 1531.40234375, "completions/mean_terminated_length": 527.8965454101562, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6694546385593582, "frac_reward_zero_std": 0.875, "grad_norm": 20.93845553695336, "learning_rate": 3.6848800828280303e-07, "loss": 0.0184, "num_tokens": 1880201642.0, "reward": 0.017578125, "reward_std": 0.04125870764255524, "rewards/accuracy_reward/mean": 0.017578125, "rewards/accuracy_reward/std": 0.13154059648513794, "step": 1961 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.779296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1986.0, "completions/mean_length": 1715.7109375, "completions/mean_terminated_length": 542.4071044921875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.669796022872749, "frac_reward_zero_std": 0.65625, "grad_norm": 62.756249492685946, "learning_rate": 3.67997576242134e-07, "loss": 0.0076, "num_tokens": 1881154006.0, "reward": 0.076171875, "reward_std": 0.12758280336856842, "rewards/accuracy_reward/mean": 0.0786290317773819, "rewards/accuracy_reward/std": 0.26943063735961914, "step": 1962 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1891.0, "completions/mean_length": 1635.578125, "completions/mean_terminated_length": 601.6986083984375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6701374071861398, "frac_reward_zero_std": 0.75, "grad_norm": 38.742672223165776, "learning_rate": 3.67507402520617e-07, "loss": -0.003, "num_tokens": 1882066158.0, "reward": 0.05078125, "reward_std": 0.08746850490570068, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 1963 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1976.0, "completions/mean_length": 1483.611328125, "completions/mean_terminated_length": 631.4951171875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6704787914995306, "frac_reward_zero_std": 0.78125, "grad_norm": 30.07118084137147, "learning_rate": 3.6701748781396367e-07, "loss": 0.0196, "num_tokens": 1882903863.0, "reward": 0.0390625, "reward_std": 0.07477159798145294, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 1964 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.712890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2002.0, "completions/mean_length": 1649.021484375, "completions/mean_terminated_length": 658.3605346679688, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.6708201758129214, "frac_reward_zero_std": 0.75, "grad_norm": 20.828769637008634, "learning_rate": 3.6652783281751873e-07, "loss": 0.004, "num_tokens": 1883833170.0, "reward": 0.064453125, "reward_std": 0.08736887574195862, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 1965 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.65234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2024.0, "completions/mean_length": 1582.546875, "completions/mean_terminated_length": 709.1685180664062, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6711615601263122, "frac_reward_zero_std": 0.78125, "grad_norm": 15.844985662898475, "learning_rate": 3.6603843822625734e-07, "loss": 0.0017, "num_tokens": 1884714458.0, "reward": 0.03125, "reward_std": 0.07147008180618286, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 1966 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.662109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2007.0, "completions/mean_length": 1587.951171875, "completions/mean_terminated_length": 686.4682006835938, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.671502944439703, "frac_reward_zero_std": 0.71875, "grad_norm": 29.01153872742641, "learning_rate": 3.6554930473478595e-07, "loss": 0.0079, "num_tokens": 1885606705.0, "reward": 0.076171875, "reward_std": 0.10592307895421982, "rewards/accuracy_reward/mean": 0.076171875, "rewards/accuracy_reward/std": 0.26553234457969666, "step": 1967 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.705078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1915.0, "completions/mean_length": 1640.8828125, "completions/mean_terminated_length": 667.576171875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6718443287530937, "frac_reward_zero_std": 0.75, "grad_norm": 146.75221062076855, "learning_rate": 3.650604330373398e-07, "loss": 0.0227, "num_tokens": 1886521445.0, "reward": 0.080078125, "reward_std": 0.10586882382631302, "rewards/accuracy_reward/mean": 0.080078125, "rewards/accuracy_reward/std": 0.271679550409317, "step": 1968 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.630859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1888.0, "completions/mean_length": 1540.52734375, "completions/mean_terminated_length": 673.2592163085938, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6721857130664846, "frac_reward_zero_std": 0.75, "grad_norm": 55.559349819124726, "learning_rate": 3.6457182382778315e-07, "loss": 0.0091, "num_tokens": 1887388739.0, "reward": 0.04296875, "reward_std": 0.08966471254825592, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 1969 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.673828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2022.0, "completions/mean_length": 1617.51953125, "completions/mean_terminated_length": 728.20361328125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.6725270973798754, "frac_reward_zero_std": 0.6875, "grad_norm": 129.9211284555465, "learning_rate": 3.6408347779960734e-07, "loss": -0.0075, "num_tokens": 1888302141.0, "reward": 0.05859375, "reward_std": 0.12026290595531464, "rewards/accuracy_reward/mean": 0.05859375, "rewards/accuracy_reward/std": 0.23509246110916138, "step": 1970 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.599609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1949.0, "completions/mean_length": 1481.630859375, "completions/mean_terminated_length": 633.4585571289062, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6728684816932662, "frac_reward_zero_std": 0.8125, "grad_norm": 15.519506172147983, "learning_rate": 3.6359539564593036e-07, "loss": 0.0032, "num_tokens": 1889130528.0, "reward": 0.0234375, "reward_std": 0.06068410724401474, "rewards/accuracy_reward/mean": 0.0234375, "rewards/accuracy_reward/std": 0.15143637359142303, "step": 1971 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1535.412109375, "completions/mean_terminated_length": 681.0989990234375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.673209866006657, "frac_reward_zero_std": 0.59375, "grad_norm": 14.521777875029994, "learning_rate": 3.631075780594952e-07, "loss": 0.0346, "num_tokens": 1889998867.0, "reward": 0.08203125, "reward_std": 0.14277496933937073, "rewards/accuracy_reward/mean": 0.08203125, "rewards/accuracy_reward/std": 0.2746807038784027, "step": 1972 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.669921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1867.0, "completions/mean_length": 1576.8359375, "completions/mean_terminated_length": 620.5680541992188, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6735512503200478, "frac_reward_zero_std": 0.625, "grad_norm": 15.804738422435477, "learning_rate": 3.626200257326697e-07, "loss": 0.0366, "num_tokens": 1890885487.0, "reward": 0.07421875, "reward_std": 0.1401873528957367, "rewards/accuracy_reward/mean": 0.07421875, "rewards/accuracy_reward/std": 0.2623828947544098, "step": 1973 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1911.0, "completions/mean_length": 1619.83984375, "completions/mean_terminated_length": 773.4767456054688, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.6738926346334386, "frac_reward_zero_std": 0.9375, "grad_norm": 27.112510943008317, "learning_rate": 3.621327393574458e-07, "loss": -0.0022, "num_tokens": 1891801165.0, "reward": 0.025390625, "reward_std": 0.031097229570150375, "rewards/accuracy_reward/mean": 0.025390625, "rewards/accuracy_reward/std": 0.15746226906776428, "step": 1974 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.72265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2044.0, "completions/mean_length": 1662.388671875, "completions/mean_terminated_length": 657.6267700195312, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.6742340189468294, "frac_reward_zero_std": 0.90625, "grad_norm": 4.470861371573744, "learning_rate": 3.616457196254367e-07, "loss": 0.0007, "num_tokens": 1892735172.0, "reward": 0.046875, "reward_std": 0.042583562433719635, "rewards/accuracy_reward/mean": 0.046875, "rewards/accuracy_reward/std": 0.21157780289649963, "step": 1975 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2003.0, "completions/mean_length": 1582.046875, "completions/mean_terminated_length": 628.2678833007812, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.6745754032602201, "frac_reward_zero_std": 0.8125, "grad_norm": 22.935143982029984, "learning_rate": 3.6115896722787833e-07, "loss": 0.0102, "num_tokens": 1893629500.0, "reward": 0.04296875, "reward_std": 0.07944431155920029, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 1976 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.634765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1994.0, "completions/mean_length": 1529.396484375, "completions/mean_terminated_length": 628.0802001953125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.674916787573611, "frac_reward_zero_std": 0.75, "grad_norm": 64.98743825334853, "learning_rate": 3.606724828556265e-07, "loss": 0.0063, "num_tokens": 1894489015.0, "reward": 0.06640625, "reward_std": 0.09739763289690018, "rewards/accuracy_reward/mean": 0.06640625, "rewards/accuracy_reward/std": 0.2492343932390213, "step": 1977 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1963.0, "completions/mean_length": 1602.3828125, "completions/mean_terminated_length": 721.5115966796875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.6752581718870018, "frac_reward_zero_std": 0.6875, "grad_norm": 55.22697989556693, "learning_rate": 3.6018626719915646e-07, "loss": 0.021, "num_tokens": 1895383963.0, "reward": 0.08984375, "reward_std": 0.12478742748498917, "rewards/accuracy_reward/mean": 0.08984375, "rewards/accuracy_reward/std": 0.2862374484539032, "step": 1978 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.62890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1926.0, "completions/mean_length": 1562.806640625, "completions/mean_terminated_length": 740.5316162109375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6755995562003926, "frac_reward_zero_std": 0.71875, "grad_norm": 29.070425650999866, "learning_rate": 3.5970032094856305e-07, "loss": 0.0118, "num_tokens": 1896253576.0, "reward": 0.05078125, "reward_std": 0.0999661535024643, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 1979 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.64453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1926.0, "completions/mean_length": 1531.51953125, "completions/mean_terminated_length": 595.0439453125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6759409405137834, "frac_reward_zero_std": 0.78125, "grad_norm": 3.2688991150374105, "learning_rate": 3.5921464479355744e-07, "loss": 0.0079, "num_tokens": 1897112402.0, "reward": 0.041015625, "reward_std": 0.07654774188995361, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 1980 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.72265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2029.0, "completions/mean_length": 1678.470703125, "completions/mean_terminated_length": 716.0211181640625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.6762823248271742, "frac_reward_zero_std": 0.53125, "grad_norm": 64.57568753126111, "learning_rate": 3.5872923942346875e-07, "loss": 0.0121, "num_tokens": 1898062099.0, "reward": 0.099609375, "reward_std": 0.17435188591480255, "rewards/accuracy_reward/mean": 0.099609375, "rewards/accuracy_reward/std": 0.29977133870124817, "step": 1981 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.697265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1901.0, "completions/mean_length": 1627.890625, "completions/mean_terminated_length": 660.2838745117188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.676623709140565, "frac_reward_zero_std": 0.625, "grad_norm": 20.200998787146453, "learning_rate": 3.582441055272406e-07, "loss": 0.0165, "num_tokens": 1898968203.0, "reward": 0.103515625, "reward_std": 0.14486464858055115, "rewards/accuracy_reward/mean": 0.103515625, "rewards/accuracy_reward/std": 0.30492907762527466, "step": 1982 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.619140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1950.0, "completions/mean_length": 1532.29296875, "completions/mean_terminated_length": 693.9384765625, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.6769650934539558, "frac_reward_zero_std": 0.84375, "grad_norm": 13.775472228414001, "learning_rate": 3.577592437934321e-07, "loss": -0.0007, "num_tokens": 1899837425.0, "reward": 0.044921875, "reward_std": 0.05913964658975601, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 1983 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.642578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2018.0, "completions/mean_length": 1550.4765625, "completions/mean_terminated_length": 656.0218505859375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6773064777673465, "frac_reward_zero_std": 0.71875, "grad_norm": 10.017215052974752, "learning_rate": 3.572746549102159e-07, "loss": 0.0143, "num_tokens": 1900705989.0, "reward": 0.03515625, "reward_std": 0.0893455371260643, "rewards/accuracy_reward/mean": 0.03629032149910927, "rewards/accuracy_reward/std": 0.1872003972530365, "step": 1984 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2041.0, "completions/mean_length": 1498.275390625, "completions/mean_terminated_length": 668.299072265625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6776478620807374, "frac_reward_zero_std": 0.59375, "grad_norm": 51.00286265154916, "learning_rate": 3.567903395653767e-07, "loss": 0.0157, "num_tokens": 1901548962.0, "reward": 0.083984375, "reward_std": 0.15112704038619995, "rewards/accuracy_reward/mean": 0.083984375, "rewards/accuracy_reward/std": 0.2776356339454651, "step": 1985 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.67578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1970.0, "completions/mean_length": 1620.279296875, "completions/mean_terminated_length": 729.0963745117188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6779892463941282, "frac_reward_zero_std": 0.8125, "grad_norm": 13.174357043836892, "learning_rate": 3.563062984463123e-07, "loss": 0.0074, "num_tokens": 1902459681.0, "reward": 0.05078125, "reward_std": 0.06794346868991852, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 1986 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.654296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1771.0, "completions/mean_length": 1522.228515625, "completions/mean_terminated_length": 527.124267578125, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.678330630707519, "frac_reward_zero_std": 0.84375, "grad_norm": 2.9424894820761374, "learning_rate": 3.5582253224003e-07, "loss": 0.0014, "num_tokens": 1903320342.0, "reward": 0.0390625, "reward_std": 0.05864075943827629, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 1987 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.716796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2015.0, "completions/mean_length": 1648.923828125, "completions/mean_terminated_length": 638.8482666015625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.6786720150209098, "frac_reward_zero_std": 0.875, "grad_norm": 17.174071424605156, "learning_rate": 3.553390416331478e-07, "loss": 0.0098, "num_tokens": 1904241327.0, "reward": 0.056640625, "reward_std": 0.03889618441462517, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 1988 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.63671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1998.0, "completions/mean_length": 1570.732421875, "completions/mean_terminated_length": 734.231201171875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.6790133993343006, "frac_reward_zero_std": 0.625, "grad_norm": 24.953005660323637, "learning_rate": 3.5485582731189176e-07, "loss": 0.017, "num_tokens": 1905123446.0, "reward": 0.109375, "reward_std": 0.1497160941362381, "rewards/accuracy_reward/mean": 0.109375, "rewards/accuracy_reward/std": 0.31241437792778015, "step": 1989 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.662109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1942.0, "completions/mean_length": 1568.1015625, "completions/mean_terminated_length": 627.7225341796875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6793547836476914, "frac_reward_zero_std": 0.6875, "grad_norm": 30.543565226885953, "learning_rate": 3.5437288996209704e-07, "loss": 0.0324, "num_tokens": 1906009642.0, "reward": 0.064453125, "reward_std": 0.11309494823217392, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 1990 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.66796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1963.0, "completions/mean_length": 1604.623046875, "completions/mean_terminated_length": 712.6529541015625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.6796961679610822, "frac_reward_zero_std": 0.875, "grad_norm": 2.3511192851781164, "learning_rate": 3.5389023026920384e-07, "loss": 0.0102, "num_tokens": 1906911961.0, "reward": 0.015625, "reward_std": 0.040274329483509064, "rewards/accuracy_reward/mean": 0.015625, "rewards/accuracy_reward/std": 0.12414088100194931, "step": 1991 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.658203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2033.0, "completions/mean_length": 1588.296875, "completions/mean_terminated_length": 703.0399780273438, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.6800375522744729, "frac_reward_zero_std": 0.78125, "grad_norm": 34.97873395477791, "learning_rate": 3.534078489182598e-07, "loss": 0.0227, "num_tokens": 1907795793.0, "reward": 0.0546875, "reward_std": 0.08698301017284393, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 1992 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.666015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1821.0, "completions/mean_length": 1587.224609375, "completions/mean_terminated_length": 668.368408203125, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.6803789365878637, "frac_reward_zero_std": 0.6875, "grad_norm": 37.86623879337408, "learning_rate": 3.5292574659391716e-07, "loss": 0.0416, "num_tokens": 1908688276.0, "reward": 0.09375, "reward_std": 0.1290597766637802, "rewards/accuracy_reward/mean": 0.09375, "rewards/accuracy_reward/std": 0.29176566004753113, "step": 1993 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1908.0, "completions/mean_length": 1562.111328125, "completions/mean_terminated_length": 665.9166870117188, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6807203209012546, "frac_reward_zero_std": 0.71875, "grad_norm": 49.96421973042858, "learning_rate": 3.524439239804318e-07, "loss": 0.0169, "num_tokens": 1909570397.0, "reward": 0.07421875, "reward_std": 0.10958899557590485, "rewards/accuracy_reward/mean": 0.07421875, "rewards/accuracy_reward/std": 0.2623828947544098, "step": 1994 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.62890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1984.0, "completions/mean_length": 1536.361328125, "completions/mean_terminated_length": 669.2684326171875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6810617052146454, "frac_reward_zero_std": 0.78125, "grad_norm": 6.824549636955396, "learning_rate": 3.519623817616629e-07, "loss": 0.0186, "num_tokens": 1910431174.0, "reward": 0.037109375, "reward_std": 0.07014618813991547, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 1995 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.611328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1977.0, "completions/mean_length": 1531.34375, "completions/mean_terminated_length": 718.7135620117188, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6814030895280362, "frac_reward_zero_std": 0.625, "grad_norm": 14.873700612979453, "learning_rate": 3.514811206210716e-07, "loss": 0.006, "num_tokens": 1911296166.0, "reward": 0.1015625, "reward_std": 0.13380376994609833, "rewards/accuracy_reward/mean": 0.1015625, "rewards/accuracy_reward/std": 0.30236753821372986, "step": 1996 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.615234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1926.0, "completions/mean_length": 1512.935546875, "completions/mean_terminated_length": 657.3756103515625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.681744473841427, "frac_reward_zero_std": 0.6875, "grad_norm": 27.926973330135663, "learning_rate": 3.5100014124171995e-07, "loss": 0.0165, "num_tokens": 1912146453.0, "reward": 0.08203125, "reward_std": 0.11630409955978394, "rewards/accuracy_reward/mean": 0.08203125, "rewards/accuracy_reward/std": 0.2746807038784027, "step": 1997 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.615234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1978.0, "completions/mean_length": 1537.765625, "completions/mean_terminated_length": 721.9086303710938, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6820858581548178, "frac_reward_zero_std": 0.71875, "grad_norm": 31.5191943322163, "learning_rate": 3.505194443062701e-07, "loss": 0.0202, "num_tokens": 1913002589.0, "reward": 0.0859375, "reward_std": 0.10394641757011414, "rewards/accuracy_reward/mean": 0.0859375, "rewards/accuracy_reward/std": 0.28054583072662354, "step": 1998 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.650390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1994.0, "completions/mean_length": 1558.443359375, "completions/mean_terminated_length": 647.7039184570312, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.6824272424682086, "frac_reward_zero_std": 0.71875, "grad_norm": 34.526846391398045, "learning_rate": 3.5003903049698356e-07, "loss": -0.0004, "num_tokens": 1913880752.0, "reward": 0.078125, "reward_std": 0.12146230041980743, "rewards/accuracy_reward/mean": 0.078125, "rewards/accuracy_reward/std": 0.26863065361976624, "step": 1999 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.630859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2020.0, "completions/mean_length": 1549.30078125, "completions/mean_terminated_length": 697.3174438476562, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.6827686267815993, "frac_reward_zero_std": 0.78125, "grad_norm": 29.881643525349762, "learning_rate": 3.4955890049572e-07, "loss": 0.0055, "num_tokens": 1914757514.0, "reward": 0.05078125, "reward_std": 0.09186187386512756, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 2000 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.677734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2009.0, "completions/mean_length": 1586.501953125, "completions/mean_terminated_length": 615.9575805664062, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.6831100110949901, "frac_reward_zero_std": 0.71875, "grad_norm": 24.18606316733675, "learning_rate": 3.490790549839359e-07, "loss": 0.0248, "num_tokens": 1915653979.0, "reward": 0.052734375, "reward_std": 0.09820909798145294, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 2001 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1800.0, "completions/mean_length": 1625.578125, "completions/mean_terminated_length": 625.1052856445312, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.683451395408381, "frac_reward_zero_std": 0.78125, "grad_norm": 6.494348631377462, "learning_rate": 3.4859949464268456e-07, "loss": 0.028, "num_tokens": 1916566595.0, "reward": 0.04296875, "reward_std": 0.08026791363954544, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 2002 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.603515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2031.0, "completions/mean_length": 1519.470703125, "completions/mean_terminated_length": 714.9605712890625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6837927797217718, "frac_reward_zero_std": 0.625, "grad_norm": 89.97144049459428, "learning_rate": 3.481202201526136e-07, "loss": -0.0001, "num_tokens": 1917424100.0, "reward": 0.107421875, "reward_std": 0.15409593284130096, "rewards/accuracy_reward/mean": 0.107421875, "rewards/accuracy_reward/std": 0.30995169281959534, "step": 2003 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.658203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1996.0, "completions/mean_length": 1587.55859375, "completions/mean_terminated_length": 700.8800048828125, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.6841341640351626, "frac_reward_zero_std": 0.8125, "grad_norm": 48.80334493609286, "learning_rate": 3.4764123219396613e-07, "loss": -0.0004, "num_tokens": 1918316898.0, "reward": 0.033203125, "reward_std": 0.06041031330823898, "rewards/accuracy_reward/mean": 0.033203125, "rewards/accuracy_reward/std": 0.17934183776378632, "step": 2004 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.728515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1864.0, "completions/mean_length": 1670.064453125, "completions/mean_terminated_length": 656.3093872070312, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6844755483485534, "frac_reward_zero_std": 0.75, "grad_norm": 32.940967126170634, "learning_rate": 3.471625314465773e-07, "loss": 0.0161, "num_tokens": 1919263267.0, "reward": 0.05859375, "reward_std": 0.0801548957824707, "rewards/accuracy_reward/mean": 0.060483869165182114, "rewards/accuracy_reward/std": 0.2386218160390854, "step": 2005 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.677734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2048.0, "completions/mean_length": 1596.724609375, "completions/mean_terminated_length": 647.6787719726562, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.6848169326619442, "frac_reward_zero_std": 0.625, "grad_norm": 48.946077928328464, "learning_rate": 3.466841185898756e-07, "loss": -0.0058, "num_tokens": 1920157862.0, "reward": 0.08203125, "reward_std": 0.13401183485984802, "rewards/accuracy_reward/mean": 0.08203125, "rewards/accuracy_reward/std": 0.2746807038784027, "step": 2006 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2023.0, "completions/mean_length": 1599.759765625, "completions/mean_terminated_length": 681.9345092773438, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.685158316975335, "frac_reward_zero_std": 0.6875, "grad_norm": 29.165821932788443, "learning_rate": 3.4620599430288077e-07, "loss": 0.0291, "num_tokens": 1921060331.0, "reward": 0.056640625, "reward_std": 0.11093294620513916, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 2007 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.591796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2023.0, "completions/mean_length": 1464.46875, "completions/mean_terminated_length": 618.488037109375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6854997012887258, "frac_reward_zero_std": 0.6875, "grad_norm": 17.291498444943254, "learning_rate": 3.457281592642026e-07, "loss": 0.0237, "num_tokens": 1921886155.0, "reward": 0.0546875, "reward_std": 0.11206455528736115, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 2008 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1887.0, "completions/mean_length": 1624.48046875, "completions/mean_terminated_length": 621.76318359375, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.6858410856021165, "frac_reward_zero_std": 0.78125, "grad_norm": 7.507376339519271, "learning_rate": 3.452506141520406e-07, "loss": 0.0118, "num_tokens": 1922802657.0, "reward": 0.025390625, "reward_std": 0.06755761802196503, "rewards/accuracy_reward/mean": 0.025390625, "rewards/accuracy_reward/std": 0.15746226906776428, "step": 2009 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.65625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2048.0, "completions/mean_length": 1577.201171875, "completions/mean_terminated_length": 678.4034423828125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6861824699155074, "frac_reward_zero_std": 0.59375, "grad_norm": 33.88163295969299, "learning_rate": 3.4477335964418237e-07, "loss": 0.0148, "num_tokens": 1923690936.0, "reward": 0.064453125, "reward_std": 0.1340966522693634, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 2010 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2015.0, "completions/mean_length": 1609.0859375, "completions/mean_terminated_length": 741.465087890625, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.6865238542288982, "frac_reward_zero_std": 0.6875, "grad_norm": 11.860884419984826, "learning_rate": 3.442963964180039e-07, "loss": 0.0159, "num_tokens": 1924595060.0, "reward": 0.05078125, "reward_std": 0.10354605317115784, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 2011 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.607421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2037.0, "completions/mean_length": 1507.072265625, "completions/mean_terminated_length": 670.1144409179688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.686865238542289, "frac_reward_zero_std": 0.8125, "grad_norm": 15.888842520814444, "learning_rate": 3.4381972515046675e-07, "loss": 0.0236, "num_tokens": 1925439961.0, "reward": 0.0234375, "reward_std": 0.06068410724401474, "rewards/accuracy_reward/mean": 0.0234375, "rewards/accuracy_reward/std": 0.15143637359142303, "step": 2012 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1983.0, "completions/mean_length": 1661.232421875, "completions/mean_terminated_length": 710.3311157226562, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.6872066228556798, "frac_reward_zero_std": 0.6875, "grad_norm": 178.3392350435034, "learning_rate": 3.4334334651811895e-07, "loss": 0.012, "num_tokens": 1926376144.0, "reward": 0.046875, "reward_std": 0.10739278793334961, "rewards/accuracy_reward/mean": 0.046875, "rewards/accuracy_reward/std": 0.21157780289649963, "step": 2013 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.658203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1911.0, "completions/mean_length": 1596.5546875, "completions/mean_terminated_length": 727.451416015625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6875480071690706, "frac_reward_zero_std": 0.75, "grad_norm": 65.18044455919652, "learning_rate": 3.428672611970931e-07, "loss": 0.0073, "num_tokens": 1927272012.0, "reward": 0.072265625, "reward_std": 0.10530310869216919, "rewards/accuracy_reward/mean": 0.072265625, "rewards/accuracy_reward/std": 0.2591804563999176, "step": 2014 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.60546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2016.0, "completions/mean_length": 1495.142578125, "completions/mean_terminated_length": 646.697998046875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6878893914824614, "frac_reward_zero_std": 0.75, "grad_norm": 17.67961907172786, "learning_rate": 3.423914698631044e-07, "loss": 0.0114, "num_tokens": 1928107317.0, "reward": 0.060546875, "reward_std": 0.10139618813991547, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.24230584502220154, "step": 2015 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.591796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2040.0, "completions/mean_length": 1415.25, "completions/mean_terminated_length": 497.9138488769531, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6882307757958522, "frac_reward_zero_std": 0.71875, "grad_norm": 20.53518517947413, "learning_rate": 3.4191597319145246e-07, "loss": 0.0242, "num_tokens": 1928913653.0, "reward": 0.0546875, "reward_std": 0.1024416983127594, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 2016 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.623046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2033.0, "completions/mean_length": 1531.533203125, "completions/mean_terminated_length": 677.8911743164062, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.6885721601092429, "frac_reward_zero_std": 0.65625, "grad_norm": 31.31642597136823, "learning_rate": 3.414407718570172e-07, "loss": 0.0224, "num_tokens": 1929776550.0, "reward": 0.078125, "reward_std": 0.12185370922088623, "rewards/accuracy_reward/mean": 0.078125, "rewards/accuracy_reward/std": 0.26863065361976624, "step": 2017 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.580078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2031.0, "completions/mean_length": 1452.70703125, "completions/mean_terminated_length": 630.3720703125, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.6889135444226338, "frac_reward_zero_std": 0.6875, "grad_norm": 83.63178004188556, "learning_rate": 3.4096586653426053e-07, "loss": 0.0256, "num_tokens": 1930594768.0, "reward": 0.0703125, "reward_std": 0.12042921781539917, "rewards/accuracy_reward/mean": 0.0703125, "rewards/accuracy_reward/std": 0.25592297315597534, "step": 2018 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.69140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.0, "completions/mean_length": 1610.703125, "completions/mean_terminated_length": 630.9367065429688, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.6892549287360246, "frac_reward_zero_std": 0.875, "grad_norm": 2.770057470535581, "learning_rate": 3.404912578972232e-07, "loss": 0.0109, "num_tokens": 1931506168.0, "reward": 0.056640625, "reward_std": 0.05629178136587143, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 2019 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2001.0, "completions/mean_length": 1529.65234375, "completions/mean_terminated_length": 636.3297729492188, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.6895963130494154, "frac_reward_zero_std": 0.78125, "grad_norm": 13.243987160739314, "learning_rate": 3.400169466195256e-07, "loss": -0.0135, "num_tokens": 1932357990.0, "reward": 0.041015625, "reward_std": 0.08279120922088623, "rewards/accuracy_reward/mean": 0.04233871027827263, "rewards/accuracy_reward/std": 0.2015640139579773, "step": 2020 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.587890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1987.0, "completions/mean_length": 1455.865234375, "completions/mean_terminated_length": 611.4265747070312, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6899376973628062, "frac_reward_zero_std": 0.625, "grad_norm": 46.3221544949549, "learning_rate": 3.395429333743663e-07, "loss": -0.0008, "num_tokens": 1933184945.0, "reward": 0.080078125, "reward_std": 0.13341790437698364, "rewards/accuracy_reward/mean": 0.080078125, "rewards/accuracy_reward/std": 0.271679550409317, "step": 2021 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.669921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1907.0, "completions/mean_length": 1583.625, "completions/mean_terminated_length": 641.1361083984375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.690279081676197, "frac_reward_zero_std": 0.75, "grad_norm": 25.388885630312487, "learning_rate": 3.3906921883451957e-07, "loss": -0.0025, "num_tokens": 1934067089.0, "reward": 0.044921875, "reward_std": 0.080595001578331, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 2022 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.62890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2001.0, "completions/mean_length": 1545.099609375, "completions/mean_terminated_length": 692.8157958984375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6906204659895878, "frac_reward_zero_std": 0.8125, "grad_norm": 6.585441697110707, "learning_rate": 3.3859580367233695e-07, "loss": 0.0133, "num_tokens": 1934924996.0, "reward": 0.0390625, "reward_std": 0.07465953379869461, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 2023 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.576171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2045.0, "completions/mean_length": 1437.66015625, "completions/mean_terminated_length": 607.9354858398438, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6909618503029786, "frac_reward_zero_std": 0.6875, "grad_norm": 149.44881492523203, "learning_rate": 3.3812268855974475e-07, "loss": 0.0265, "num_tokens": 1935739894.0, "reward": 0.0546875, "reward_std": 0.10767117142677307, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 2024 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.66796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1824.0, "completions/mean_length": 1563.283203125, "completions/mean_terminated_length": 588.1470947265625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.6913032346163693, "frac_reward_zero_std": 0.6875, "grad_norm": 40.85489343336783, "learning_rate": 3.376498741682433e-07, "loss": -0.0124, "num_tokens": 1936622983.0, "reward": 0.05078125, "reward_std": 0.09594620764255524, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 2025 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.673828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1962.0, "completions/mean_length": 1583.158203125, "completions/mean_terminated_length": 622.8563232421875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.6916446189297601, "frac_reward_zero_std": 0.6875, "grad_norm": 43.54316556914517, "learning_rate": 3.3717736116890585e-07, "loss": 0.0196, "num_tokens": 1937508840.0, "reward": 0.068359375, "reward_std": 0.12263333797454834, "rewards/accuracy_reward/mean": 0.068359375, "rewards/accuracy_reward/std": 0.25260838866233826, "step": 2026 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2042.0, "completions/mean_length": 1568.451171875, "completions/mean_terminated_length": 683.9500122070312, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.691986003243151, "frac_reward_zero_std": 0.6875, "grad_norm": 21.1005545821204, "learning_rate": 3.3670515023237866e-07, "loss": 0.0021, "num_tokens": 1938383375.0, "reward": 0.099609375, "reward_std": 0.12997153401374817, "rewards/accuracy_reward/mean": 0.099609375, "rewards/accuracy_reward/std": 0.29977133870124817, "step": 2027 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.69140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1962.0, "completions/mean_length": 1596.982421875, "completions/mean_terminated_length": 586.4746704101562, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6923273875565418, "frac_reward_zero_std": 0.75, "grad_norm": 23.851763842090207, "learning_rate": 3.362332420288786e-07, "loss": 0.0077, "num_tokens": 1939287910.0, "reward": 0.04296875, "reward_std": 0.08263044059276581, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 2028 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.66015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1992.0, "completions/mean_length": 1582.94140625, "completions/mean_terminated_length": 679.9022827148438, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6926687718699326, "frac_reward_zero_std": 0.71875, "grad_norm": 83.74372184694154, "learning_rate": 3.3576163722819273e-07, "loss": 0.0214, "num_tokens": 1940188472.0, "reward": 0.080078125, "reward_std": 0.11623015999794006, "rewards/accuracy_reward/mean": 0.080078125, "rewards/accuracy_reward/std": 0.271679550409317, "step": 2029 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.65625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2039.0, "completions/mean_length": 1554.3671875, "completions/mean_terminated_length": 611.977294921875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6930101561833234, "frac_reward_zero_std": 0.65625, "grad_norm": 44.28437423992409, "learning_rate": 3.3529033649967843e-07, "loss": 0.0419, "num_tokens": 1941059412.0, "reward": 0.09765625, "reward_std": 0.14422638714313507, "rewards/accuracy_reward/mean": 0.09765625, "rewards/accuracy_reward/std": 0.29713961482048035, "step": 2030 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.587890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1937.0, "completions/mean_length": 1461.978515625, "completions/mean_terminated_length": 625.9953002929688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6933515404967142, "frac_reward_zero_std": 0.75, "grad_norm": 38.05531967898502, "learning_rate": 3.3481934051226024e-07, "loss": -0.0007, "num_tokens": 1941891193.0, "reward": 0.03515625, "reward_std": 0.08274346590042114, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 2031 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.59375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1823.0, "completions/mean_length": 1467.142578125, "completions/mean_terminated_length": 618.1971435546875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.693692924810105, "frac_reward_zero_std": 0.71875, "grad_norm": 11.89901083510781, "learning_rate": 3.3434864993443123e-07, "loss": 0.0379, "num_tokens": 1942707986.0, "reward": 0.0859375, "reward_std": 0.10056467354297638, "rewards/accuracy_reward/mean": 0.0859375, "rewards/accuracy_reward/std": 0.28054583072662354, "step": 2032 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.67578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2008.0, "completions/mean_length": 1607.7578125, "completions/mean_terminated_length": 690.4156494140625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6940343091234957, "frac_reward_zero_std": 0.75, "grad_norm": 19.748402295301776, "learning_rate": 3.338782654342506e-07, "loss": 0.008, "num_tokens": 1943618646.0, "reward": 0.041015625, "reward_std": 0.08406735956668854, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 2033 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.599609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1997.0, "completions/mean_length": 1510.283203125, "completions/mean_terminated_length": 705.01953125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6943756934368865, "frac_reward_zero_std": 0.78125, "grad_norm": 21.367762265059742, "learning_rate": 3.334081876793427e-07, "loss": 0.0031, "num_tokens": 1944464039.0, "reward": 0.08203125, "reward_std": 0.08670367300510406, "rewards/accuracy_reward/mean": 0.08203125, "rewards/accuracy_reward/std": 0.2746807038784027, "step": 2034 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.630859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1974.0, "completions/mean_length": 1514.12109375, "completions/mean_terminated_length": 601.724853515625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6947170777502774, "frac_reward_zero_std": 0.625, "grad_norm": 30.023164434329964, "learning_rate": 3.3293841733689745e-07, "loss": 0.0167, "num_tokens": 1945317813.0, "reward": 0.111328125, "reward_std": 0.15493440628051758, "rewards/accuracy_reward/mean": 0.111328125, "rewards/accuracy_reward/std": 0.31484565138816833, "step": 2035 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.638671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1956.0, "completions/mean_length": 1545.2421875, "completions/mean_terminated_length": 656.5838012695312, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6950584620636682, "frac_reward_zero_std": 0.78125, "grad_norm": 20.21608847885543, "learning_rate": 3.324689550736674e-07, "loss": -0.0085, "num_tokens": 1946183697.0, "reward": 0.029296875, "reward_std": 0.06288585811853409, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 2036 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.634765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1997.0, "completions/mean_length": 1536.029296875, "completions/mean_terminated_length": 646.2406616210938, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.695399846377059, "frac_reward_zero_std": 0.84375, "grad_norm": 77.23223590948912, "learning_rate": 3.3199980155596895e-07, "loss": -0.0007, "num_tokens": 1947054112.0, "reward": 0.0390625, "reward_std": 0.060571081936359406, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 2037 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.603515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1947.0, "completions/mean_length": 1528.05078125, "completions/mean_terminated_length": 736.6009521484375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6957412306904498, "frac_reward_zero_std": 0.75, "grad_norm": 25.969533425263116, "learning_rate": 3.315309574496792e-07, "loss": 0.006, "num_tokens": 1947912778.0, "reward": 0.060546875, "reward_std": 0.08537881821393967, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.24230584502220154, "step": 2038 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.69140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2041.0, "completions/mean_length": 1624.8046875, "completions/mean_terminated_length": 676.9683837890625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.6960826150038406, "frac_reward_zero_std": 0.71875, "grad_norm": 64.35573317486414, "learning_rate": 3.310624234202369e-07, "loss": 0.0179, "num_tokens": 1948825462.0, "reward": 0.07421875, "reward_std": 0.11824101209640503, "rewards/accuracy_reward/mean": 0.07661290466785431, "rewards/accuracy_reward/std": 0.2662447690963745, "step": 2039 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.712890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2000.0, "completions/mean_length": 1667.25390625, "completions/mean_terminated_length": 721.8639526367188, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.6964239993172314, "frac_reward_zero_std": 0.78125, "grad_norm": 122.23248740945724, "learning_rate": 3.305942001326404e-07, "loss": 0.0037, "num_tokens": 1949756584.0, "reward": 0.0546875, "reward_std": 0.08604401350021362, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 2040 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.634765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1938.0, "completions/mean_length": 1511.685546875, "completions/mean_terminated_length": 579.5882568359375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.6967653836306221, "frac_reward_zero_std": 0.6875, "grad_norm": 92.51768673437242, "learning_rate": 3.3012628825144685e-07, "loss": 0.0043, "num_tokens": 1950612055.0, "reward": 0.0546875, "reward_std": 0.11214477568864822, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 2041 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.701171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1958.0, "completions/mean_length": 1650.369140625, "completions/mean_terminated_length": 717.3660278320312, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6971067679440129, "frac_reward_zero_std": 0.78125, "grad_norm": 32.92817516635831, "learning_rate": 3.296586884407717e-07, "loss": 0.0117, "num_tokens": 1951538308.0, "reward": 0.08203125, "reward_std": 0.08963288366794586, "rewards/accuracy_reward/mean": 0.08203125, "rewards/accuracy_reward/std": 0.2746807038784027, "step": 2042 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.689453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2006.0, "completions/mean_length": 1628.591796875, "completions/mean_terminated_length": 697.4528198242188, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.6974481522574038, "frac_reward_zero_std": 0.84375, "grad_norm": 43.31814393751971, "learning_rate": 3.2919140136428727e-07, "loss": 0.0172, "num_tokens": 1952445811.0, "reward": 0.048828125, "reward_std": 0.05919293314218521, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 2043 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.66015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1914.0, "completions/mean_length": 1565.32421875, "completions/mean_terminated_length": 627.712646484375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6977895365707946, "frac_reward_zero_std": 0.78125, "grad_norm": 143.95312631651663, "learning_rate": 3.287244276852223e-07, "loss": 0.0188, "num_tokens": 1953321673.0, "reward": 0.03125, "reward_std": 0.07152433693408966, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 2044 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2021.0, "completions/mean_length": 1559.3046875, "completions/mean_terminated_length": 657.933349609375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6981309208841854, "frac_reward_zero_std": 0.78125, "grad_norm": 55.58020166957072, "learning_rate": 3.282577680663604e-07, "loss": 0.0021, "num_tokens": 1954201573.0, "reward": 0.068359375, "reward_std": 0.09020963311195374, "rewards/accuracy_reward/mean": 0.068359375, "rewards/accuracy_reward/std": 0.25260838866233826, "step": 2045 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.62890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1836.0, "completions/mean_length": 1533.001953125, "completions/mean_terminated_length": 660.2158203125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6984723051975762, "frac_reward_zero_std": 0.78125, "grad_norm": 43.689061727235156, "learning_rate": 3.277914231700393e-07, "loss": 0.0113, "num_tokens": 1955066854.0, "reward": 0.04296875, "reward_std": 0.06964729726314545, "rewards/accuracy_reward/mean": 0.04435483738780022, "rewards/accuracy_reward/std": 0.2060900777578354, "step": 2046 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2001.0, "completions/mean_length": 1533.478515625, "completions/mean_terminated_length": 675.9427490234375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.698813689510967, "frac_reward_zero_std": 0.78125, "grad_norm": 41.02430527797403, "learning_rate": 3.273253936581506e-07, "loss": 0.0009, "num_tokens": 1955925179.0, "reward": 0.0390625, "reward_std": 0.08422812819480896, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 2047 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.623046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1516.904296875, "completions/mean_terminated_length": 639.0828857421875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.6991550738243578, "frac_reward_zero_std": 0.78125, "grad_norm": 9.82404224891299, "learning_rate": 3.2685968019213784e-07, "loss": 0.0174, "num_tokens": 1956776074.0, "reward": 0.03515625, "reward_std": 0.07443207502365112, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 2048 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.69140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1983.0, "completions/mean_length": 1634.53515625, "completions/mean_terminated_length": 708.16455078125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6994964581377485, "frac_reward_zero_std": 0.90625, "grad_norm": 16.615725340369927, "learning_rate": 3.2639428343299623e-07, "loss": 0.004, "num_tokens": 1957687068.0, "reward": 0.015625, "reward_std": 0.03344620764255524, "rewards/accuracy_reward/mean": 0.015625, "rewards/accuracy_reward/std": 0.12414088100194931, "step": 2049 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.673828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2036.0, "completions/mean_length": 1597.525390625, "completions/mean_terminated_length": 666.9042358398438, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.6998378424511393, "frac_reward_zero_std": 0.59375, "grad_norm": 92.29997608715432, "learning_rate": 3.259292040412711e-07, "loss": 0.0364, "num_tokens": 1958581017.0, "reward": 0.115234375, "reward_std": 0.16516265273094177, "rewards/accuracy_reward/mean": 0.115234375, "rewards/accuracy_reward/std": 0.3196168541908264, "step": 2050 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.705078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2007.0, "completions/mean_length": 1634.677734375, "completions/mean_terminated_length": 646.5364379882812, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7001792267645301, "frac_reward_zero_std": 0.75, "grad_norm": 41.52746029897798, "learning_rate": 3.2546444267705786e-07, "loss": -0.0008, "num_tokens": 1959491588.0, "reward": 0.03515625, "reward_std": 0.08214399218559265, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 2051 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.685546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1991.0, "completions/mean_length": 1603.701171875, "completions/mean_terminated_length": 635.3975219726562, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.700520611077921, "frac_reward_zero_std": 0.78125, "grad_norm": 32.7995522967481, "learning_rate": 3.250000000000001e-07, "loss": 0.0142, "num_tokens": 1960388971.0, "reward": 0.048828125, "reward_std": 0.08184495568275452, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 2052 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.685546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1946.0, "completions/mean_length": 1625.08984375, "completions/mean_terminated_length": 703.0932006835938, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7008619953913118, "frac_reward_zero_std": 0.8125, "grad_norm": 13.156427950796258, "learning_rate": 3.245358766692891e-07, "loss": 0.0152, "num_tokens": 1961297337.0, "reward": 0.060546875, "reward_std": 0.07069281488656998, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 2053 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.685546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1939.0, "completions/mean_length": 1620.775390625, "completions/mean_terminated_length": 689.3726806640625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7012033797047026, "frac_reward_zero_std": 0.84375, "grad_norm": 22.684303239790186, "learning_rate": 3.2407207334366347e-07, "loss": 0.0053, "num_tokens": 1962204070.0, "reward": 0.021484375, "reward_std": 0.054354868829250336, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 2054 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.63671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1993.0, "completions/mean_length": 1535.599609375, "completions/mean_terminated_length": 637.521484375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7015447640180934, "frac_reward_zero_std": 0.75, "grad_norm": 16.571154698087064, "learning_rate": 3.2360859068140666e-07, "loss": 0.0289, "num_tokens": 1963069001.0, "reward": 0.048828125, "reward_std": 0.09363143146038055, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 2055 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2044.0, "completions/mean_length": 1539.08203125, "completions/mean_terminated_length": 631.8804321289062, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7018861483314842, "frac_reward_zero_std": 0.75, "grad_norm": 35.260464535352945, "learning_rate": 3.2314542934034813e-07, "loss": 0.01, "num_tokens": 1963928771.0, "reward": 0.0703125, "reward_std": 0.09970590472221375, "rewards/accuracy_reward/mean": 0.0703125, "rewards/accuracy_reward/std": 0.25592297315597534, "step": 2056 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1954.0, "completions/mean_length": 1594.025390625, "completions/mean_terminated_length": 664.4583740234375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7022275326448749, "frac_reward_zero_std": 0.78125, "grad_norm": 12.448077060752237, "learning_rate": 3.2268258997786015e-07, "loss": 0.0045, "num_tokens": 1964820656.0, "reward": 0.02734375, "reward_std": 0.07284127175807953, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 2057 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.701171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1855.0, "completions/mean_length": 1622.609375, "completions/mean_terminated_length": 624.7777709960938, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7025689169582657, "frac_reward_zero_std": 0.65625, "grad_norm": 46.664486883601015, "learning_rate": 3.2222007325085885e-07, "loss": 0.0144, "num_tokens": 1965731752.0, "reward": 0.103515625, "reward_std": 0.1336519569158554, "rewards/accuracy_reward/mean": 0.10685484111309052, "rewards/accuracy_reward/std": 0.3092404901981354, "step": 2058 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.638671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1992.0, "completions/mean_length": 1548.703125, "completions/mean_terminated_length": 666.1621704101562, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.7029103012716565, "frac_reward_zero_std": 0.65625, "grad_norm": 130.9995498795912, "learning_rate": 3.217578798158022e-07, "loss": 0.0167, "num_tokens": 1966597760.0, "reward": 0.10546875, "reward_std": 0.13027715682983398, "rewards/accuracy_reward/mean": 0.10546875, "rewards/accuracy_reward/std": 0.3074568510055542, "step": 2059 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.65234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2045.0, "completions/mean_length": 1572.05859375, "completions/mean_terminated_length": 679.0, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7032516855850474, "frac_reward_zero_std": 0.65625, "grad_norm": 121.6637216076443, "learning_rate": 3.2129601032868884e-07, "loss": 0.0046, "num_tokens": 1967482206.0, "reward": 0.1015625, "reward_std": 0.11228963732719421, "rewards/accuracy_reward/mean": 0.1015625, "rewards/accuracy_reward/std": 0.30236753821372986, "step": 2060 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.767578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2040.0, "completions/mean_length": 1748.810546875, "completions/mean_terminated_length": 761.2521362304688, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.7035930698984382, "frac_reward_zero_std": 0.6875, "grad_norm": 45.72616905030464, "learning_rate": 3.2083446544505847e-07, "loss": 0.0007, "num_tokens": 1968467661.0, "reward": 0.07421875, "reward_std": 0.11989058554172516, "rewards/accuracy_reward/mean": 0.07421875, "rewards/accuracy_reward/std": 0.2623828947544098, "step": 2061 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.70703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2019.0, "completions/mean_length": 1657.28515625, "completions/mean_terminated_length": 714.3600463867188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.703934454211829, "frac_reward_zero_std": 0.75, "grad_norm": 123.5932995574792, "learning_rate": 3.203732458199893e-07, "loss": -0.0061, "num_tokens": 1969389807.0, "reward": 0.1015625, "reward_std": 0.10079116374254227, "rewards/accuracy_reward/mean": 0.1015625, "rewards/accuracy_reward/std": 0.30236753821372986, "step": 2062 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.693359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1946.0, "completions/mean_length": 1629.236328125, "completions/mean_terminated_length": 682.350341796875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7042758385252198, "frac_reward_zero_std": 0.65625, "grad_norm": 45.09570648828714, "learning_rate": 3.199123521080985e-07, "loss": 0.0048, "num_tokens": 1970304040.0, "reward": 0.091796875, "reward_std": 0.13095036149024963, "rewards/accuracy_reward/mean": 0.091796875, "rewards/accuracy_reward/std": 0.289021372795105, "step": 2063 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.64453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2037.0, "completions/mean_length": 1542.720703125, "completions/mean_terminated_length": 626.8516845703125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7046172228386106, "frac_reward_zero_std": 0.75, "grad_norm": 19.692904742401137, "learning_rate": 3.194517849635404e-07, "loss": 0.0114, "num_tokens": 1971179097.0, "reward": 0.046875, "reward_std": 0.08780898153781891, "rewards/accuracy_reward/mean": 0.04838709533214569, "rewards/accuracy_reward/std": 0.21479946374893188, "step": 2064 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.65234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2010.0, "completions/mean_length": 1546.52734375, "completions/mean_terminated_length": 605.5618286132812, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7049586071520013, "frac_reward_zero_std": 0.78125, "grad_norm": 3.801723405728734, "learning_rate": 3.1899154504000544e-07, "loss": 0.0257, "num_tokens": 1972046631.0, "reward": 0.064453125, "reward_std": 0.0971047431230545, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 2065 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.646484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2027.0, "completions/mean_length": 1573.609375, "completions/mean_terminated_length": 706.077392578125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.7052999914653921, "frac_reward_zero_std": 0.75, "grad_norm": 40.66839924548039, "learning_rate": 3.185316329907204e-07, "loss": -0.0019, "num_tokens": 1972931407.0, "reward": 0.048828125, "reward_std": 0.07493096590042114, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 2066 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.662109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2040.0, "completions/mean_length": 1541.583984375, "completions/mean_terminated_length": 549.2427368164062, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7056413757787829, "frac_reward_zero_std": 0.875, "grad_norm": 7.918111794638964, "learning_rate": 3.1807204946844613e-07, "loss": 0.0027, "num_tokens": 1973797530.0, "reward": 0.03515625, "reward_std": 0.05530741065740585, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 2067 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.693359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1965.0, "completions/mean_length": 1648.8359375, "completions/mean_terminated_length": 746.5732421875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.7059827600921738, "frac_reward_zero_std": 0.78125, "grad_norm": 8.01388166767705, "learning_rate": 3.176127951254775e-07, "loss": 0.0405, "num_tokens": 1974723142.0, "reward": 0.064453125, "reward_std": 0.08879335969686508, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 2068 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.650390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1973.0, "completions/mean_length": 1560.298828125, "completions/mean_terminated_length": 653.0111694335938, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7063241444055646, "frac_reward_zero_std": 0.8125, "grad_norm": 11.955420253573317, "learning_rate": 3.1715387061364187e-07, "loss": 0.0168, "num_tokens": 1975603567.0, "reward": 0.033203125, "reward_std": 0.06570751965045929, "rewards/accuracy_reward/mean": 0.033203125, "rewards/accuracy_reward/std": 0.17934183776378632, "step": 2069 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.669921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1988.0, "completions/mean_length": 1580.921875, "completions/mean_terminated_length": 632.94677734375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7066655287189554, "frac_reward_zero_std": 0.625, "grad_norm": 29.018306524252484, "learning_rate": 3.1669527658429914e-07, "loss": 0.0129, "num_tokens": 1976486311.0, "reward": 0.0703125, "reward_std": 0.13132822513580322, "rewards/accuracy_reward/mean": 0.07258064299821854, "rewards/accuracy_reward/std": 0.25970885157585144, "step": 2070 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.603515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2017.0, "completions/mean_length": 1482.923828125, "completions/mean_terminated_length": 622.7832641601562, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.7070069130323462, "frac_reward_zero_std": 0.65625, "grad_norm": 33.92418337490979, "learning_rate": 3.162370136883389e-07, "loss": 0.0008, "num_tokens": 1977326848.0, "reward": 0.056640625, "reward_std": 0.11383409798145294, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 2071 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1961.0, "completions/mean_length": 1578.40625, "completions/mean_terminated_length": 650.1395263671875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.707348297345737, "frac_reward_zero_std": 0.78125, "grad_norm": 15.304359804523731, "learning_rate": 3.157790825761818e-07, "loss": -0.006, "num_tokens": 1978210096.0, "reward": 0.021484375, "reward_std": 0.06519509106874466, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 2072 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.67578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1693.0, "completions/mean_length": 1582.2734375, "completions/mean_terminated_length": 611.5421752929688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7076896816591277, "frac_reward_zero_std": 0.6875, "grad_norm": 21.017632317170996, "learning_rate": 3.1532148389777766e-07, "loss": 0.0082, "num_tokens": 1979090476.0, "reward": 0.08203125, "reward_std": 0.11835404485464096, "rewards/accuracy_reward/mean": 0.08203125, "rewards/accuracy_reward/std": 0.2746807038784027, "step": 2073 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1980.0, "completions/mean_length": 1604.4609375, "completions/mean_terminated_length": 663.2926635742188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7080310659725185, "frac_reward_zero_std": 0.75, "grad_norm": 112.35664168327071, "learning_rate": 3.148642183026037e-07, "loss": 0.019, "num_tokens": 1979990728.0, "reward": 0.044921875, "reward_std": 0.0830162987112999, "rewards/accuracy_reward/mean": 0.0463709682226181, "rewards/accuracy_reward/std": 0.21049949526786804, "step": 2074 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.638671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1969.0, "completions/mean_length": 1541.19140625, "completions/mean_terminated_length": 645.3729858398438, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7083724502859093, "frac_reward_zero_std": 0.65625, "grad_norm": 36.25662558441624, "learning_rate": 3.14407286439665e-07, "loss": 0.0175, "num_tokens": 1980857546.0, "reward": 0.09765625, "reward_std": 0.14319565892219543, "rewards/accuracy_reward/mean": 0.09765625, "rewards/accuracy_reward/std": 0.29713961482048035, "step": 2075 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.62109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1908.0, "completions/mean_length": 1499.466796875, "completions/mean_terminated_length": 600.32470703125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7087138345993002, "frac_reward_zero_std": 0.8125, "grad_norm": 39.591695047263805, "learning_rate": 3.1395068895749275e-07, "loss": -0.0001, "num_tokens": 1981707641.0, "reward": 0.033203125, "reward_std": 0.06695909798145294, "rewards/accuracy_reward/mean": 0.03427419438958168, "rewards/accuracy_reward/std": 0.18211629986763, "step": 2076 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.68359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1880.0, "completions/mean_length": 1615.04296875, "completions/mean_terminated_length": 679.6419677734375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.709055218912691, "frac_reward_zero_std": 0.75, "grad_norm": 10.45142760747069, "learning_rate": 3.134944265041436e-07, "loss": 0.0277, "num_tokens": 1982611551.0, "reward": 0.07421875, "reward_std": 0.09948080778121948, "rewards/accuracy_reward/mean": 0.07421875, "rewards/accuracy_reward/std": 0.2623828947544098, "step": 2077 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.642578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2028.0, "completions/mean_length": 1547.646484375, "completions/mean_terminated_length": 648.1038208007812, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.7093966032260818, "frac_reward_zero_std": 0.59375, "grad_norm": 35.088840288631104, "learning_rate": 3.1303849972719834e-07, "loss": 0.0087, "num_tokens": 1983474970.0, "reward": 0.0625, "reward_std": 0.13896197080612183, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.2422981858253479, "step": 2078 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.69921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2019.0, "completions/mean_length": 1632.234375, "completions/mean_terminated_length": 665.7142944335938, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7097379875394726, "frac_reward_zero_std": 0.875, "grad_norm": 94.36120064906882, "learning_rate": 3.1258290927376187e-07, "loss": -0.0052, "num_tokens": 1984390338.0, "reward": 0.060546875, "reward_std": 0.04726085811853409, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 2079 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.62890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1512.21484375, "completions/mean_terminated_length": 604.2000122070312, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7100793718528634, "frac_reward_zero_std": 0.84375, "grad_norm": 28.74341934206451, "learning_rate": 3.121276557904616e-07, "loss": 0.007, "num_tokens": 1985241840.0, "reward": 0.02734375, "reward_std": 0.05402229726314545, "rewards/accuracy_reward/mean": 0.02822580561041832, "rewards/accuracy_reward/std": 0.1657845675945282, "step": 2080 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.638671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1786.0, "completions/mean_length": 1529.40625, "completions/mean_terminated_length": 612.7567749023438, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7104207561662541, "frac_reward_zero_std": 0.6875, "grad_norm": 108.9552564520055, "learning_rate": 3.1167273992344646e-07, "loss": 0.0045, "num_tokens": 1986103360.0, "reward": 0.04296875, "reward_std": 0.10668125748634338, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 2081 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.705078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1601.0, "completions/mean_length": 1608.630859375, "completions/mean_terminated_length": 558.2185668945312, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7107621404796449, "frac_reward_zero_std": 0.875, "grad_norm": 62.19652651673116, "learning_rate": 3.112181623183866e-07, "loss": 0.0124, "num_tokens": 1987003587.0, "reward": 0.025390625, "reward_std": 0.04620979726314545, "rewards/accuracy_reward/mean": 0.025390625, "rewards/accuracy_reward/std": 0.15746226906776428, "step": 2082 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.572265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1883.0, "completions/mean_length": 1425.36328125, "completions/mean_terminated_length": 592.337890625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7111035247930357, "frac_reward_zero_std": 0.78125, "grad_norm": 11.071100894793616, "learning_rate": 3.1076392362047117e-07, "loss": -0.0032, "num_tokens": 1987806013.0, "reward": 0.07421875, "reward_std": 0.07662828266620636, "rewards/accuracy_reward/mean": 0.07421875, "rewards/accuracy_reward/std": 0.2623828947544098, "step": 2083 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.63671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2034.0, "completions/mean_length": 1555.46484375, "completions/mean_terminated_length": 692.2042846679688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7114449091064265, "frac_reward_zero_std": 0.625, "grad_norm": 56.09849120397067, "learning_rate": 3.1031002447440945e-07, "loss": 0.0071, "num_tokens": 1988671899.0, "reward": 0.083984375, "reward_std": 0.13721179962158203, "rewards/accuracy_reward/mean": 0.083984375, "rewards/accuracy_reward/std": 0.2776356339454651, "step": 2084 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.62109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2031.0, "completions/mean_length": 1529.162109375, "completions/mean_terminated_length": 678.6958618164062, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7117862934198174, "frac_reward_zero_std": 0.75, "grad_norm": 43.497110095292925, "learning_rate": 3.0985646552442794e-07, "loss": 0.0071, "num_tokens": 1989531278.0, "reward": 0.0703125, "reward_std": 0.10056467354297638, "rewards/accuracy_reward/mean": 0.0703125, "rewards/accuracy_reward/std": 0.25592297315597534, "step": 2085 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.626953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1983.0, "completions/mean_length": 1516.572265625, "completions/mean_terminated_length": 623.4398193359375, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.7121276777332082, "frac_reward_zero_std": 0.78125, "grad_norm": 64.54069808962709, "learning_rate": 3.0940324741427103e-07, "loss": 0.0091, "num_tokens": 1990382259.0, "reward": 0.025390625, "reward_std": 0.06755761057138443, "rewards/accuracy_reward/mean": 0.025390625, "rewards/accuracy_reward/std": 0.15746226906776428, "step": 2086 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.67578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1960.0, "completions/mean_length": 1597.0703125, "completions/mean_terminated_length": 657.1807250976562, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.712469062046599, "frac_reward_zero_std": 0.625, "grad_norm": 54.81338224920416, "learning_rate": 3.089503707871983e-07, "loss": 0.0008, "num_tokens": 1991281719.0, "reward": 0.072265625, "reward_std": 0.13717760145664215, "rewards/accuracy_reward/mean": 0.072265625, "rewards/accuracy_reward/std": 0.2591804563999176, "step": 2087 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2031.0, "completions/mean_length": 1611.068359375, "completions/mean_terminated_length": 683.9207153320312, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7128104463599898, "frac_reward_zero_std": 0.6875, "grad_norm": 13.602775862880675, "learning_rate": 3.084978362859859e-07, "loss": 0.0178, "num_tokens": 1992189898.0, "reward": 0.060546875, "reward_std": 0.1127297431230545, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 2088 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.669921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1939.0, "completions/mean_length": 1595.0546875, "completions/mean_terminated_length": 675.7633056640625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7131518306733805, "frac_reward_zero_std": 0.78125, "grad_norm": 22.659904632380997, "learning_rate": 3.080456445529237e-07, "loss": 0.0169, "num_tokens": 1993083910.0, "reward": 0.0625, "reward_std": 0.08141306042671204, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.2422981858253479, "step": 2089 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.697265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1827.0, "completions/mean_length": 1598.7734375, "completions/mean_terminated_length": 564.1032104492188, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7134932149867713, "frac_reward_zero_std": 0.6875, "grad_norm": 24.79468584601153, "learning_rate": 3.075937962298147e-07, "loss": 0.0227, "num_tokens": 1993975330.0, "reward": 0.056640625, "reward_std": 0.1127830445766449, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 2090 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.66796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1883.0, "completions/mean_length": 1560.078125, "completions/mean_terminated_length": 578.494140625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7138345993001621, "frac_reward_zero_std": 0.6875, "grad_norm": 43.48332757833183, "learning_rate": 3.0714229195797545e-07, "loss": 0.0052, "num_tokens": 1994849098.0, "reward": 0.146484375, "reward_std": 0.14131072163581848, "rewards/accuracy_reward/mean": 0.15120968222618103, "rewards/accuracy_reward/std": 0.35861483216285706, "step": 2091 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.673828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1807.0, "completions/mean_length": 1594.40234375, "completions/mean_terminated_length": 657.329345703125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.714175983613553, "frac_reward_zero_std": 0.78125, "grad_norm": 27.578527355911092, "learning_rate": 3.066911323782333e-07, "loss": 0.0042, "num_tokens": 1995745496.0, "reward": 0.046875, "reward_std": 0.08584587275981903, "rewards/accuracy_reward/mean": 0.046875, "rewards/accuracy_reward/std": 0.21157780289649963, "step": 2092 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.705078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2009.0, "completions/mean_length": 1644.96875, "completions/mean_terminated_length": 681.4304809570312, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.7145173679269438, "frac_reward_zero_std": 0.71875, "grad_norm": 68.83237109358673, "learning_rate": 3.062403181309271e-07, "loss": -0.0, "num_tokens": 1996660504.0, "reward": 0.068359375, "reward_std": 0.11270354688167572, "rewards/accuracy_reward/mean": 0.068359375, "rewards/accuracy_reward/std": 0.25260838866233826, "step": 2093 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.65234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1922.0, "completions/mean_length": 1532.369140625, "completions/mean_terminated_length": 564.8370971679688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7148587522403346, "frac_reward_zero_std": 0.78125, "grad_norm": 5.666846989593071, "learning_rate": 3.057898498559049e-07, "loss": 0.0153, "num_tokens": 1997519349.0, "reward": 0.0546875, "reward_std": 0.08698301017284393, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 2094 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2018.0, "completions/mean_length": 1642.39453125, "completions/mean_terminated_length": 644.8243408203125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7152001365537254, "frac_reward_zero_std": 0.78125, "grad_norm": 25.381641768321064, "learning_rate": 3.053397281925244e-07, "loss": 0.017, "num_tokens": 1998431839.0, "reward": 0.029296875, "reward_std": 0.07328138500452042, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 2095 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.662109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1961.0, "completions/mean_length": 1567.37109375, "completions/mean_terminated_length": 625.5606689453125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7155415208671162, "frac_reward_zero_std": 0.75, "grad_norm": 127.87539587907906, "learning_rate": 3.0488995377965064e-07, "loss": -0.0152, "num_tokens": 1999310413.0, "reward": 0.06640625, "reward_std": 0.10053187608718872, "rewards/accuracy_reward/mean": 0.06640625, "rewards/accuracy_reward/std": 0.2492343932390213, "step": 2096 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.732421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1958.0, "completions/mean_length": 1679.142578125, "completions/mean_terminated_length": 669.496337890625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7158829051805069, "frac_reward_zero_std": 0.71875, "grad_norm": 49.93796341915726, "learning_rate": 3.0444052725565614e-07, "loss": 0.0054, "num_tokens": 2000241174.0, "reward": 0.0703125, "reward_std": 0.11086063086986542, "rewards/accuracy_reward/mean": 0.0703125, "rewards/accuracy_reward/std": 0.25592297315597534, "step": 2097 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.66015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1839.0, "completions/mean_length": 1534.58203125, "completions/mean_terminated_length": 537.2528686523438, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.7162242894938977, "frac_reward_zero_std": 0.75, "grad_norm": 51.8223202132066, "learning_rate": 3.0399144925841993e-07, "loss": 0.0161, "num_tokens": 2001110800.0, "reward": 0.044921875, "reward_std": 0.08423367142677307, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 2098 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.626953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1991.0, "completions/mean_length": 1518.9453125, "completions/mean_terminated_length": 630.0576171875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7165656738072885, "frac_reward_zero_std": 0.78125, "grad_norm": 33.55706020272989, "learning_rate": 3.0354272042532573e-07, "loss": -0.0083, "num_tokens": 2001961508.0, "reward": 0.064453125, "reward_std": 0.09500166773796082, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 2099 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1859.0, "completions/mean_length": 1484.490234375, "completions/mean_terminated_length": 633.7009887695312, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.7169070581206793, "frac_reward_zero_std": 0.65625, "grad_norm": 25.69656031424839, "learning_rate": 3.030943413932624e-07, "loss": 0.0466, "num_tokens": 2002790287.0, "reward": 0.09765625, "reward_std": 0.1418900489807129, "rewards/accuracy_reward/mean": 0.10080645233392715, "rewards/accuracy_reward/std": 0.30137622356414795, "step": 2100 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.68359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1935.0, "completions/mean_length": 1584.056640625, "completions/mean_terminated_length": 582.0493774414062, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7172484424340702, "frac_reward_zero_std": 0.59375, "grad_norm": 61.92127933297947, "learning_rate": 3.0264631279862183e-07, "loss": 0.0166, "num_tokens": 2003682508.0, "reward": 0.091796875, "reward_std": 0.15636999905109406, "rewards/accuracy_reward/mean": 0.091796875, "rewards/accuracy_reward/std": 0.289021372795105, "step": 2101 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.5546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1997.0, "completions/mean_length": 1417.962890625, "completions/mean_terminated_length": 633.1798095703125, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.717589826747461, "frac_reward_zero_std": 0.65625, "grad_norm": 23.53358520748768, "learning_rate": 3.021986352772985e-07, "loss": 0.0105, "num_tokens": 2004492057.0, "reward": 0.0859375, "reward_std": 0.12737178802490234, "rewards/accuracy_reward/mean": 0.0859375, "rewards/accuracy_reward/std": 0.28054583072662354, "step": 2102 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.634765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2007.0, "completions/mean_length": 1517.890625, "completions/mean_terminated_length": 596.5775756835938, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7179312110608518, "frac_reward_zero_std": 0.65625, "grad_norm": 26.339287537232327, "learning_rate": 3.0175130946468894e-07, "loss": 0.0332, "num_tokens": 2005347009.0, "reward": 0.0625, "reward_std": 0.12478180229663849, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.2422981858253479, "step": 2103 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2034.0, "completions/mean_length": 1594.30078125, "completions/mean_terminated_length": 665.297607421875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7182725953742426, "frac_reward_zero_std": 0.71875, "grad_norm": 70.86283084805004, "learning_rate": 3.013043359956903e-07, "loss": 0.0103, "num_tokens": 2006240123.0, "reward": 0.064453125, "reward_std": 0.10288181900978088, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 2104 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.67578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1973.0, "completions/mean_length": 1588.08984375, "completions/mean_terminated_length": 629.48193359375, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.7186139796876334, "frac_reward_zero_std": 0.78125, "grad_norm": 112.06093086486976, "learning_rate": 3.008577155046997e-07, "loss": 0.0232, "num_tokens": 2007133433.0, "reward": 0.044921875, "reward_std": 0.08204546570777893, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 2105 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.638671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1884.0, "completions/mean_length": 1558.294921875, "completions/mean_terminated_length": 692.7081298828125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7189553640010241, "frac_reward_zero_std": 0.84375, "grad_norm": 2.2251411968431123, "learning_rate": 3.0041144862561307e-07, "loss": 0.0019, "num_tokens": 2008006944.0, "reward": 0.041015625, "reward_std": 0.05907990783452988, "rewards/accuracy_reward/mean": 0.04233871027827263, "rewards/accuracy_reward/std": 0.2015640139579773, "step": 2106 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.587890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1992.0, "completions/mean_length": 1472.7109375, "completions/mean_terminated_length": 652.0379638671875, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.7192967483144149, "frac_reward_zero_std": 0.6875, "grad_norm": 11.372619449927114, "learning_rate": 2.9996553599182487e-07, "loss": -0.0022, "num_tokens": 2008841164.0, "reward": 0.072265625, "reward_std": 0.11669091880321503, "rewards/accuracy_reward/mean": 0.072265625, "rewards/accuracy_reward/std": 0.2591804563999176, "step": 2107 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.5390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2031.0, "completions/mean_length": 1400.58203125, "completions/mean_terminated_length": 643.4321899414062, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.7196381326278057, "frac_reward_zero_std": 0.71875, "grad_norm": 17.94512173951747, "learning_rate": 2.995199782362264e-07, "loss": -0.0018, "num_tokens": 2009633462.0, "reward": 0.05859375, "reward_std": 0.10013246536254883, "rewards/accuracy_reward/mean": 0.05859375, "rewards/accuracy_reward/std": 0.23509246110916138, "step": 2108 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1925.0, "completions/mean_length": 1502.3671875, "completions/mean_terminated_length": 622.6734619140625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.7199795169411966, "frac_reward_zero_std": 0.65625, "grad_norm": 25.795837495466298, "learning_rate": 2.9907477599120537e-07, "loss": 0.0188, "num_tokens": 2010475474.0, "reward": 0.083984375, "reward_std": 0.12340269237756729, "rewards/accuracy_reward/mean": 0.083984375, "rewards/accuracy_reward/std": 0.2776356339454651, "step": 2109 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.626953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1999.0, "completions/mean_length": 1535.6015625, "completions/mean_terminated_length": 674.4502563476562, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7203209012545874, "frac_reward_zero_std": 0.875, "grad_norm": 1.9266647590356358, "learning_rate": 2.98629929888645e-07, "loss": 0.0114, "num_tokens": 2011344630.0, "reward": 0.025390625, "reward_std": 0.044359706342220306, "rewards/accuracy_reward/mean": 0.025390625, "rewards/accuracy_reward/std": 0.15746226906776428, "step": 2110 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1953.0, "completions/mean_length": 1552.669921875, "completions/mean_terminated_length": 639.3611450195312, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.7206622855679782, "frac_reward_zero_std": 0.90625, "grad_norm": 0.3908546761184745, "learning_rate": 2.981854405599228e-07, "loss": 0.0024, "num_tokens": 2012221277.0, "reward": 0.0078125, "reward_std": 0.02629890665411949, "rewards/accuracy_reward/mean": 0.0078125, "rewards/accuracy_reward/std": 0.08812850713729858, "step": 2111 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.55078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1952.0, "completions/mean_length": 1423.806640625, "completions/mean_terminated_length": 658.4912719726562, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.721003669881369, "frac_reward_zero_std": 0.65625, "grad_norm": 116.44842924076096, "learning_rate": 2.9774130863591035e-07, "loss": 0.0453, "num_tokens": 2013026202.0, "reward": 0.1015625, "reward_std": 0.12988576292991638, "rewards/accuracy_reward/mean": 0.1015625, "rewards/accuracy_reward/std": 0.30236753821372986, "step": 2112 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2043.0, "completions/mean_length": 1547.572265625, "completions/mean_terminated_length": 624.5610961914062, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7213450541947598, "frac_reward_zero_std": 0.59375, "grad_norm": 31.148788561009, "learning_rate": 2.9729753474697157e-07, "loss": 0.0333, "num_tokens": 2013892783.0, "reward": 0.1015625, "reward_std": 0.1591646373271942, "rewards/accuracy_reward/mean": 0.1015625, "rewards/accuracy_reward/std": 0.30236753821372986, "step": 2113 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.59375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1931.0, "completions/mean_length": 1509.60546875, "completions/mean_terminated_length": 722.72119140625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.7216864385081505, "frac_reward_zero_std": 0.65625, "grad_norm": 34.16179392993526, "learning_rate": 2.9685411952296214e-07, "loss": 0.0261, "num_tokens": 2014736101.0, "reward": 0.046875, "reward_std": 0.11322767287492752, "rewards/accuracy_reward/mean": 0.04838709533214569, "rewards/accuracy_reward/std": 0.21479946374893188, "step": 2114 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.603515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2033.0, "completions/mean_length": 1452.013671875, "completions/mean_terminated_length": 545.1083374023438, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7220278228215413, "frac_reward_zero_std": 0.59375, "grad_norm": 6.371800954336332, "learning_rate": 2.964110635932292e-07, "loss": 0.0532, "num_tokens": 2015559372.0, "reward": 0.064453125, "reward_std": 0.13831710815429688, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 2115 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.697265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1930.0, "completions/mean_length": 1620.830078125, "completions/mean_terminated_length": 636.9613037109375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7223692071349321, "frac_reward_zero_std": 0.71875, "grad_norm": 20.994038027963132, "learning_rate": 2.959683675866093e-07, "loss": 0.0047, "num_tokens": 2016461797.0, "reward": 0.03125, "reward_std": 0.08846627175807953, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 2116 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.701171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2043.0, "completions/mean_length": 1637.13671875, "completions/mean_terminated_length": 673.0849609375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.722710591448323, "frac_reward_zero_std": 0.6875, "grad_norm": 77.60443361460818, "learning_rate": 2.955260321314287e-07, "loss": 0.0211, "num_tokens": 2017382955.0, "reward": 0.09375, "reward_std": 0.13056449592113495, "rewards/accuracy_reward/mean": 0.09677419066429138, "rewards/accuracy_reward/std": 0.2959485352039337, "step": 2117 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.740234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1756.0, "completions/mean_length": 1686.826171875, "completions/mean_terminated_length": 657.9849853515625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.7230519757617138, "frac_reward_zero_std": 0.5625, "grad_norm": 106.6455793241189, "learning_rate": 2.9508405785550144e-07, "loss": 0.0225, "num_tokens": 2018325874.0, "reward": 0.08984375, "reward_std": 0.15241675078868866, "rewards/accuracy_reward/mean": 0.08984375, "rewards/accuracy_reward/std": 0.2862374484539032, "step": 2118 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.68359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2029.0, "completions/mean_length": 1622.09765625, "completions/mean_terminated_length": 702.2098999023438, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7233933600751046, "frac_reward_zero_std": 0.78125, "grad_norm": 32.25114948899705, "learning_rate": 2.946424453861294e-07, "loss": -0.0033, "num_tokens": 2019242916.0, "reward": 0.025390625, "reward_std": 0.06849660724401474, "rewards/accuracy_reward/mean": 0.026209676638245583, "rewards/accuracy_reward/std": 0.1599196344614029, "step": 2119 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.75, "completions/max_length": 2048.0, "completions/max_terminated_length": 1946.0, "completions/mean_length": 1709.6953125, "completions/mean_terminated_length": 695.203125, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.7237347443884954, "frac_reward_zero_std": 0.8125, "grad_norm": 15.140927571795668, "learning_rate": 2.942011953501007e-07, "loss": 0.015, "num_tokens": 2020193608.0, "reward": 0.05859375, "reward_std": 0.08235109597444534, "rewards/accuracy_reward/mean": 0.05859375, "rewards/accuracy_reward/std": 0.23509246110916138, "step": 2120 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2040.0, "completions/mean_length": 1635.654296875, "completions/mean_terminated_length": 659.4144897460938, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.7240761287018862, "frac_reward_zero_std": 0.59375, "grad_norm": 78.35080327197711, "learning_rate": 2.937603083736887e-07, "loss": 0.0125, "num_tokens": 2021119735.0, "reward": 0.09375, "reward_std": 0.1529373973608017, "rewards/accuracy_reward/mean": 0.09375, "rewards/accuracy_reward/std": 0.29176566004753113, "step": 2121 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1915.0, "completions/mean_length": 1633.21484375, "completions/mean_terminated_length": 593.4109497070312, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.7244175130152769, "frac_reward_zero_std": 0.625, "grad_norm": 87.49090776121162, "learning_rate": 2.9331978508265225e-07, "loss": -0.0034, "num_tokens": 2022034181.0, "reward": 0.08203125, "reward_std": 0.12917280197143555, "rewards/accuracy_reward/mean": 0.08203125, "rewards/accuracy_reward/std": 0.2746807038784027, "step": 2122 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1845.0, "completions/mean_length": 1579.50390625, "completions/mean_terminated_length": 585.719482421875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7247588973286677, "frac_reward_zero_std": 0.84375, "grad_norm": 1.1251685996495475, "learning_rate": 2.9287962610223326e-07, "loss": 0.0159, "num_tokens": 2022921863.0, "reward": 0.029296875, "reward_std": 0.057768452912569046, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17417415976524353, "step": 2123 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1878.0, "completions/mean_length": 1631.65625, "completions/mean_terminated_length": 607.6757202148438, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7251002816420585, "frac_reward_zero_std": 0.78125, "grad_norm": 18.247549999673293, "learning_rate": 2.924398320571573e-07, "loss": -0.0052, "num_tokens": 2023833671.0, "reward": 0.048828125, "reward_std": 0.08275700360536575, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 2124 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.673828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2013.0, "completions/mean_length": 1576.05859375, "completions/mean_terminated_length": 601.08984375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7254416659554493, "frac_reward_zero_std": 0.6875, "grad_norm": 80.06744771309165, "learning_rate": 2.9200040357163114e-07, "loss": 0.0201, "num_tokens": 2024718389.0, "reward": 0.099609375, "reward_std": 0.11494503915309906, "rewards/accuracy_reward/mean": 0.099609375, "rewards/accuracy_reward/std": 0.29977133870124817, "step": 2125 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.615234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2007.0, "completions/mean_length": 1512.138671875, "completions/mean_terminated_length": 655.3045654296875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.7257830502688402, "frac_reward_zero_std": 0.71875, "grad_norm": 10.939497160811365, "learning_rate": 2.915613412693435e-07, "loss": 0.0099, "num_tokens": 2025569260.0, "reward": 0.0859375, "reward_std": 0.1209416463971138, "rewards/accuracy_reward/mean": 0.0859375, "rewards/accuracy_reward/std": 0.28054583072662354, "step": 2126 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.75390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1864.0, "completions/mean_length": 1687.203125, "completions/mean_terminated_length": 582.3651123046875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.726124434582231, "frac_reward_zero_std": 0.75, "grad_norm": 10.493811249008756, "learning_rate": 2.911226457734628e-07, "loss": 0.0181, "num_tokens": 2026520916.0, "reward": 0.041015625, "reward_std": 0.09352485090494156, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 2127 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2044.0, "completions/mean_length": 1547.080078125, "completions/mean_terminated_length": 556.8895263671875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7264658188956218, "frac_reward_zero_std": 0.78125, "grad_norm": 32.6843699245717, "learning_rate": 2.90684317706637e-07, "loss": 0.018, "num_tokens": 2027395677.0, "reward": 0.0390625, "reward_std": 0.08422909677028656, "rewards/accuracy_reward/mean": 0.04032257944345474, "rewards/accuracy_reward/std": 0.19691328704357147, "step": 2128 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1981.0, "completions/mean_length": 1513.82421875, "completions/mean_terminated_length": 593.223388671875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.7268072032090126, "frac_reward_zero_std": 0.625, "grad_norm": 11.907946947362637, "learning_rate": 2.9024635769099287e-07, "loss": 0.0107, "num_tokens": 2028245523.0, "reward": 0.05078125, "reward_std": 0.12279270589351654, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 2129 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1811.0, "completions/mean_length": 1525.5234375, "completions/mean_terminated_length": 561.844482421875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7271485875224033, "frac_reward_zero_std": 0.71875, "grad_norm": 23.061260393837998, "learning_rate": 2.8980876634813424e-07, "loss": -0.0119, "num_tokens": 2029105615.0, "reward": 0.0703125, "reward_std": 0.1009584367275238, "rewards/accuracy_reward/mean": 0.07258064299821854, "rewards/accuracy_reward/std": 0.25970885157585144, "step": 2130 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.591796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1709.0, "completions/mean_length": 1473.966796875, "completions/mean_terminated_length": 641.7559814453125, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.7274899718357941, "frac_reward_zero_std": 0.625, "grad_norm": 93.75600436920634, "learning_rate": 2.8937154429914233e-07, "loss": 0.016, "num_tokens": 2029933166.0, "reward": 0.08203125, "reward_std": 0.12538030743598938, "rewards/accuracy_reward/mean": 0.08203125, "rewards/accuracy_reward/std": 0.2746807038784027, "step": 2131 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.65234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1977.0, "completions/mean_length": 1583.759765625, "completions/mean_terminated_length": 712.6572875976562, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.7278313561491849, "frac_reward_zero_std": 0.71875, "grad_norm": 27.260480554920285, "learning_rate": 2.889346921645737e-07, "loss": 0.0091, "num_tokens": 2030823507.0, "reward": 0.04296875, "reward_std": 0.0957798957824707, "rewards/accuracy_reward/mean": 0.04435483738780022, "rewards/accuracy_reward/std": 0.2060900777578354, "step": 2132 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.646484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2005.0, "completions/mean_length": 1561.701171875, "completions/mean_terminated_length": 672.3922729492188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7281727404625757, "frac_reward_zero_std": 0.71875, "grad_norm": 21.598119000094954, "learning_rate": 2.8849821056445983e-07, "loss": 0.0095, "num_tokens": 2031697674.0, "reward": 0.037109375, "reward_std": 0.09001073241233826, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 2133 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2016.0, "completions/mean_length": 1561.9296875, "completions/mean_terminated_length": 665.9166870117188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7285141247759666, "frac_reward_zero_std": 0.875, "grad_norm": 30.932063536092194, "learning_rate": 2.880621001183069e-07, "loss": 0.0129, "num_tokens": 2032577494.0, "reward": 0.033203125, "reward_std": 0.04726085811853409, "rewards/accuracy_reward/mean": 0.033203125, "rewards/accuracy_reward/std": 0.17934183776378632, "step": 2134 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.626953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2018.0, "completions/mean_length": 1543.669921875, "completions/mean_terminated_length": 696.0785522460938, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.7288555090893574, "frac_reward_zero_std": 0.65625, "grad_norm": 24.87981394803063, "learning_rate": 2.8762636144509366e-07, "loss": 0.0294, "num_tokens": 2033440189.0, "reward": 0.123046875, "reward_std": 0.12631139159202576, "rewards/accuracy_reward/mean": 0.123046875, "rewards/accuracy_reward/std": 0.32881227135658264, "step": 2135 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.748046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2026.0, "completions/mean_length": 1689.712890625, "completions/mean_terminated_length": 625.9612426757812, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7291968934027482, "frac_reward_zero_std": 0.90625, "grad_norm": 4.826984093652752, "learning_rate": 2.871909951632716e-07, "loss": 0.0018, "num_tokens": 2034387706.0, "reward": 0.02734375, "reward_std": 0.03944835811853409, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 2136 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.66796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2040.0, "completions/mean_length": 1607.451171875, "completions/mean_terminated_length": 721.1705932617188, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.729538277716139, "frac_reward_zero_std": 0.6875, "grad_norm": 162.6916226319031, "learning_rate": 2.867560018907634e-07, "loss": 0.0144, "num_tokens": 2035295457.0, "reward": 0.03515625, "reward_std": 0.09452171623706818, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 2137 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.701171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2041.0, "completions/mean_length": 1639.021484375, "completions/mean_terminated_length": 679.3921508789062, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.7298796620295297, "frac_reward_zero_std": 0.6875, "grad_norm": 6.375931532454759, "learning_rate": 2.863213822449629e-07, "loss": 0.0109, "num_tokens": 2036213036.0, "reward": 0.0625, "reward_std": 0.11454009264707565, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.2422981858253479, "step": 2138 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1994.0, "completions/mean_length": 1677.818359375, "completions/mean_terminated_length": 654.375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7302210463429205, "frac_reward_zero_std": 0.875, "grad_norm": 24.615095473629758, "learning_rate": 2.8588713684273247e-07, "loss": 0.0112, "num_tokens": 2037148047.0, "reward": 0.017578125, "reward_std": 0.04125870764255524, "rewards/accuracy_reward/mean": 0.017578125, "rewards/accuracy_reward/std": 0.13154059648513794, "step": 2139 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1918.0, "completions/mean_length": 1597.84765625, "completions/mean_terminated_length": 642.6463012695312, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7305624306563113, "frac_reward_zero_std": 0.75, "grad_norm": 16.6170912042429, "learning_rate": 2.8545326630040436e-07, "loss": 0.0051, "num_tokens": 2038041505.0, "reward": 0.0859375, "reward_std": 0.0876617580652237, "rewards/accuracy_reward/mean": 0.0859375, "rewards/accuracy_reward/std": 0.28054583072662354, "step": 2140 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.740234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1991.0, "completions/mean_length": 1684.169921875, "completions/mean_terminated_length": 647.3909912109375, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.7309038149697021, "frac_reward_zero_std": 0.78125, "grad_norm": 123.20078432377602, "learning_rate": 2.850197712337786e-07, "loss": -0.008, "num_tokens": 2038982616.0, "reward": 0.029296875, "reward_std": 0.06288585811853409, "rewards/accuracy_reward/mean": 0.030241934582591057, "rewards/accuracy_reward/std": 0.1714252382516861, "step": 2141 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.712890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1912.0, "completions/mean_length": 1606.841796875, "completions/mean_terminated_length": 511.448974609375, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.731245199283093, "frac_reward_zero_std": 0.6875, "grad_norm": 24.665294381701703, "learning_rate": 2.84586652258122e-07, "loss": 0.02, "num_tokens": 2039876487.0, "reward": 0.0625, "reward_std": 0.11926513910293579, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.2422981858253479, "step": 2142 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.681640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1920.0, "completions/mean_length": 1611.685546875, "completions/mean_terminated_length": 677.4907836914062, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7315865835964838, "frac_reward_zero_std": 0.625, "grad_norm": 33.53266727016599, "learning_rate": 2.841539099881678e-07, "loss": -0.0006, "num_tokens": 2040777190.0, "reward": 0.05078125, "reward_std": 0.11493149399757385, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 2143 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.654296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1888.0, "completions/mean_length": 1565.595703125, "completions/mean_terminated_length": 652.5706176757812, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7319279679098746, "frac_reward_zero_std": 0.78125, "grad_norm": 23.921330018514727, "learning_rate": 2.837215450381144e-07, "loss": 0.0125, "num_tokens": 2041655671.0, "reward": 0.046875, "reward_std": 0.08490687608718872, "rewards/accuracy_reward/mean": 0.046875, "rewards/accuracy_reward/std": 0.21157780289649963, "step": 2144 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.63671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1944.0, "completions/mean_length": 1508.40625, "completions/mean_terminated_length": 562.6666870117188, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7322693522232654, "frac_reward_zero_std": 0.84375, "grad_norm": 408.78243737744873, "learning_rate": 2.832895580216249e-07, "loss": -0.01, "num_tokens": 2042504375.0, "reward": 0.0234375, "reward_std": 0.05385598540306091, "rewards/accuracy_reward/mean": 0.0234375, "rewards/accuracy_reward/std": 0.15143637359142303, "step": 2145 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.669921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1968.0, "completions/mean_length": 1591.685546875, "completions/mean_terminated_length": 665.5562133789062, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7326107365366561, "frac_reward_zero_std": 0.71875, "grad_norm": 33.833339106538666, "learning_rate": 2.828579495518256e-07, "loss": 0.0197, "num_tokens": 2043400342.0, "reward": 0.087890625, "reward_std": 0.09886875748634338, "rewards/accuracy_reward/mean": 0.087890625, "rewards/accuracy_reward/std": 0.2834126651287079, "step": 2146 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.666015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1943.0, "completions/mean_length": 1597.1171875, "completions/mean_terminated_length": 698.2924194335938, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7329521208500469, "frac_reward_zero_std": 0.71875, "grad_norm": 18.74986535595416, "learning_rate": 2.824267202413061e-07, "loss": 0.0041, "num_tokens": 2044300146.0, "reward": 0.0390625, "reward_std": 0.09655256569385529, "rewards/accuracy_reward/mean": 0.04032257944345474, "rewards/accuracy_reward/std": 0.19691328704357147, "step": 2147 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.73046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2015.0, "completions/mean_length": 1661.197265625, "completions/mean_terminated_length": 612.9058227539062, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.7332935051634377, "frac_reward_zero_std": 0.65625, "grad_norm": 54.985629609165294, "learning_rate": 2.8199587070211737e-07, "loss": -0.0134, "num_tokens": 2045238807.0, "reward": 0.068359375, "reward_std": 0.13329237699508667, "rewards/accuracy_reward/mean": 0.068359375, "rewards/accuracy_reward/std": 0.25260838866233826, "step": 2148 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.69140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1922.0, "completions/mean_length": 1609.6953125, "completions/mean_terminated_length": 627.9746704101562, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7336348894768285, "frac_reward_zero_std": 0.6875, "grad_norm": 22.758863577655593, "learning_rate": 2.815654015457715e-07, "loss": 0.0003, "num_tokens": 2046142971.0, "reward": 0.0625, "reward_std": 0.11399487406015396, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.2422981858253479, "step": 2149 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.60546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2012.0, "completions/mean_length": 1447.439453125, "completions/mean_terminated_length": 525.787109375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7339762737902193, "frac_reward_zero_std": 0.75, "grad_norm": 5.33310318371966, "learning_rate": 2.8113531338324104e-07, "loss": 0.0456, "num_tokens": 2046956908.0, "reward": 0.083984375, "reward_std": 0.09721681475639343, "rewards/accuracy_reward/mean": 0.083984375, "rewards/accuracy_reward/std": 0.2776356339454651, "step": 2150 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.59765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1981.0, "completions/mean_length": 1468.966796875, "completions/mean_terminated_length": 608.8495483398438, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.7343176581036102, "frac_reward_zero_std": 0.625, "grad_norm": 45.62410379278095, "learning_rate": 2.807056068249569e-07, "loss": -0.0047, "num_tokens": 2047781355.0, "reward": 0.107421875, "reward_std": 0.14394694566726685, "rewards/accuracy_reward/mean": 0.107421875, "rewards/accuracy_reward/std": 0.30995169281959534, "step": 2151 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.66015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2043.0, "completions/mean_length": 1566.19921875, "completions/mean_terminated_length": 630.287353515625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.734659042417001, "frac_reward_zero_std": 0.59375, "grad_norm": 4.828628081196631, "learning_rate": 2.8027628248080944e-07, "loss": 0.0532, "num_tokens": 2048657921.0, "reward": 0.1171875, "reward_std": 0.1552657186985016, "rewards/accuracy_reward/mean": 0.1171875, "rewards/accuracy_reward/std": 0.32195815443992615, "step": 2152 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.64453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2012.0, "completions/mean_length": 1564.083984375, "completions/mean_terminated_length": 686.6538696289062, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7350004267303918, "frac_reward_zero_std": 0.71875, "grad_norm": 129.09812042631734, "learning_rate": 2.7984734096014567e-07, "loss": -0.0013, "num_tokens": 2049530508.0, "reward": 0.064453125, "reward_std": 0.10791876912117004, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 2153 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1775.0, "completions/mean_length": 1630.28515625, "completions/mean_terminated_length": 677.0384521484375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7353418110437825, "frac_reward_zero_std": 0.71875, "grad_norm": 15.892844275676405, "learning_rate": 2.794187828717698e-07, "loss": 0.0005, "num_tokens": 2050438734.0, "reward": 0.044921875, "reward_std": 0.08412160724401474, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 2154 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.716796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2019.0, "completions/mean_length": 1659.203125, "completions/mean_terminated_length": 675.5379028320312, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7356831953571733, "frac_reward_zero_std": 0.75, "grad_norm": 19.962784308527507, "learning_rate": 2.789906088239419e-07, "loss": -0.0031, "num_tokens": 2051373174.0, "reward": 0.078125, "reward_std": 0.09811057895421982, "rewards/accuracy_reward/mean": 0.078125, "rewards/accuracy_reward/std": 0.26863065361976624, "step": 2155 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.685546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1899.0, "completions/mean_length": 1632.462890625, "completions/mean_terminated_length": 726.5404052734375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7360245796705641, "frac_reward_zero_std": 0.6875, "grad_norm": 21.156284615498176, "learning_rate": 2.7856281942437635e-07, "loss": 0.0033, "num_tokens": 2052285187.0, "reward": 0.0703125, "reward_std": 0.11658906936645508, "rewards/accuracy_reward/mean": 0.0703125, "rewards/accuracy_reward/std": 0.25592297315597534, "step": 2156 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.67578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2025.0, "completions/mean_length": 1616.939453125, "completions/mean_terminated_length": 718.4638061523438, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7363659639839549, "frac_reward_zero_std": 0.8125, "grad_norm": 12.253528561246215, "learning_rate": 2.781354152802422e-07, "loss": 0.0192, "num_tokens": 2053182788.0, "reward": 0.033203125, "reward_std": 0.06750432401895523, "rewards/accuracy_reward/mean": 0.033203125, "rewards/accuracy_reward/std": 0.17934183776378632, "step": 2157 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.728515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2043.0, "completions/mean_length": 1683.3671875, "completions/mean_terminated_length": 704.89208984375, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.7367073482973457, "frac_reward_zero_std": 0.75, "grad_norm": 50.28575735998116, "learning_rate": 2.777083969981611e-07, "loss": -0.0033, "num_tokens": 2054131792.0, "reward": 0.037109375, "reward_std": 0.08537977933883667, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 2158 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.619140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2022.0, "completions/mean_length": 1519.478515625, "completions/mean_terminated_length": 660.2923583984375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7370487326107366, "frac_reward_zero_std": 0.8125, "grad_norm": 35.30051746518248, "learning_rate": 2.7728176518420786e-07, "loss": 0.0037, "num_tokens": 2054984917.0, "reward": 0.0390625, "reward_std": 0.07465953379869461, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 2159 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.623046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1958.0, "completions/mean_length": 1559.234375, "completions/mean_terminated_length": 751.3782348632812, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7373901169241274, "frac_reward_zero_std": 0.78125, "grad_norm": 2.4318451217488195, "learning_rate": 2.768555204439079e-07, "loss": 0.0214, "num_tokens": 2055861325.0, "reward": 0.03515625, "reward_std": 0.07779236882925034, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 2160 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.66015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2024.0, "completions/mean_length": 1574.849609375, "completions/mean_terminated_length": 655.7413940429688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7377315012375182, "frac_reward_zero_std": 0.6875, "grad_norm": 33.94684259827365, "learning_rate": 2.764296633822379e-07, "loss": 0.0434, "num_tokens": 2056748160.0, "reward": 0.078125, "reward_std": 0.11186403036117554, "rewards/accuracy_reward/mean": 0.078125, "rewards/accuracy_reward/std": 0.26863065361976624, "step": 2161 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2010.0, "completions/mean_length": 1558.080078125, "completions/mean_terminated_length": 684.7445678710938, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.7380728855509089, "frac_reward_zero_std": 0.6875, "grad_norm": 60.527931257144765, "learning_rate": 2.7600419460362416e-07, "loss": -0.0093, "num_tokens": 2057617881.0, "reward": 0.064453125, "reward_std": 0.12121544778347015, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 2162 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.66796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1966.0, "completions/mean_length": 1565.96484375, "completions/mean_terminated_length": 596.2235107421875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7384142698642997, "frac_reward_zero_std": 0.6875, "grad_norm": 22.361342581883594, "learning_rate": 2.7557911471194167e-07, "loss": 0.0094, "num_tokens": 2058502711.0, "reward": 0.0546875, "reward_std": 0.10821780562400818, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 2163 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.626953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1916.0, "completions/mean_length": 1533.142578125, "completions/mean_terminated_length": 667.858642578125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.7387556541776905, "frac_reward_zero_std": 0.71875, "grad_norm": 45.85597883938025, "learning_rate": 2.7515442431051363e-07, "loss": 0.0137, "num_tokens": 2059363856.0, "reward": 0.037109375, "reward_std": 0.0946291983127594, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 2164 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1613.0, "completions/mean_length": 1629.923828125, "completions/mean_terminated_length": 639.743408203125, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.7390970384910813, "frac_reward_zero_std": 0.78125, "grad_norm": 2.820880402168814, "learning_rate": 2.747301240021101e-07, "loss": 0.0063, "num_tokens": 2060266537.0, "reward": 0.0390625, "reward_std": 0.06563520431518555, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 2165 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.611328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1971.0, "completions/mean_length": 1481.8671875, "completions/mean_terminated_length": 591.4170532226562, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7394384228044721, "frac_reward_zero_std": 0.78125, "grad_norm": 21.805108863433123, "learning_rate": 2.7430621438894816e-07, "loss": 0.0114, "num_tokens": 2061106533.0, "reward": 0.025390625, "reward_std": 0.06755761057138443, "rewards/accuracy_reward/mean": 0.025390625, "rewards/accuracy_reward/std": 0.15746226906776428, "step": 2166 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.63671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1988.0, "completions/mean_length": 1527.873046875, "completions/mean_terminated_length": 616.252685546875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.739779807117863, "frac_reward_zero_std": 0.75, "grad_norm": 30.656893189894657, "learning_rate": 2.7388269607268967e-07, "loss": 0.0056, "num_tokens": 2061964900.0, "reward": 0.044921875, "reward_std": 0.09369117021560669, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 2167 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.716796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1771.0, "completions/mean_length": 1640.94921875, "completions/mean_terminated_length": 611.0689697265625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7401211914312538, "frac_reward_zero_std": 0.6875, "grad_norm": 29.407697003197022, "learning_rate": 2.734595696544416e-07, "loss": 0.0145, "num_tokens": 2062891722.0, "reward": 0.052734375, "reward_std": 0.1062953919172287, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 2168 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2035.0, "completions/mean_length": 1635.716796875, "completions/mean_terminated_length": 694.8654174804688, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.7404625757446446, "frac_reward_zero_std": 0.75, "grad_norm": 45.01250673342925, "learning_rate": 2.730368357347548e-07, "loss": 0.0071, "num_tokens": 2063803785.0, "reward": 0.048828125, "reward_std": 0.09380322694778442, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 2169 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1953.0, "completions/mean_length": 1538.35546875, "completions/mean_terminated_length": 660.0319213867188, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.7408039600580353, "frac_reward_zero_std": 0.65625, "grad_norm": 11.373094852052875, "learning_rate": 2.7261449491362197e-07, "loss": 0.0227, "num_tokens": 2064671775.0, "reward": 0.115234375, "reward_std": 0.136160746216774, "rewards/accuracy_reward/mean": 0.115234375, "rewards/accuracy_reward/std": 0.3196168541908264, "step": 2170 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.666015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2029.0, "completions/mean_length": 1568.78515625, "completions/mean_terminated_length": 613.1578979492188, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.7411453443714261, "frac_reward_zero_std": 0.65625, "grad_norm": 19.92048153553588, "learning_rate": 2.721925477904794e-07, "loss": 0.0085, "num_tokens": 2065550017.0, "reward": 0.056640625, "reward_std": 0.11261671781539917, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 2171 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.62109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1764.0, "completions/mean_length": 1496.28515625, "completions/mean_terminated_length": 591.9277954101562, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.7414867286848169, "frac_reward_zero_std": 0.78125, "grad_norm": 32.187289055574006, "learning_rate": 2.717709949642034e-07, "loss": 0.0126, "num_tokens": 2066386595.0, "reward": 0.04296875, "reward_std": 0.07873040437698364, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 2172 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.58203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2009.0, "completions/mean_length": 1460.166015625, "completions/mean_terminated_length": 641.5934448242188, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7418281129982077, "frac_reward_zero_std": 0.75, "grad_norm": 78.86628871886052, "learning_rate": 2.7134983703311136e-07, "loss": 0.0348, "num_tokens": 2067211976.0, "reward": 0.05078125, "reward_std": 0.0918210819363594, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 2173 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.595703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1929.0, "completions/mean_length": 1448.177734375, "completions/mean_terminated_length": 564.3816528320312, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.7421694973115985, "frac_reward_zero_std": 0.78125, "grad_norm": 27.311016956466656, "learning_rate": 2.7092907459495973e-07, "loss": 0.0135, "num_tokens": 2068028291.0, "reward": 0.037109375, "reward_std": 0.07768725603818893, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 2174 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.728515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1867.0, "completions/mean_length": 1684.357421875, "completions/mean_terminated_length": 708.5396118164062, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.7425108816249893, "frac_reward_zero_std": 0.78125, "grad_norm": 8.541195220484141, "learning_rate": 2.7050870824694407e-07, "loss": 0.0207, "num_tokens": 2068974298.0, "reward": 0.0390625, "reward_std": 0.06948098540306091, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 2175 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.67578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1943.0, "completions/mean_length": 1619.640625, "completions/mean_terminated_length": 727.1325073242188, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7428522659383802, "frac_reward_zero_std": 0.84375, "grad_norm": 0.13536573905432692, "learning_rate": 2.700887385856974e-07, "loss": 0.0057, "num_tokens": 2069882978.0, "reward": 0.05859375, "reward_std": 0.05806133896112442, "rewards/accuracy_reward/mean": 0.060483869165182114, "rewards/accuracy_reward/std": 0.2386218160390854, "step": 2176 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1932.0, "completions/mean_length": 1648.224609375, "completions/mean_terminated_length": 626.576416015625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.743193650251771, "frac_reward_zero_std": 0.9375, "grad_norm": 10.49712994726984, "learning_rate": 2.6966916620728966e-07, "loss": -0.0, "num_tokens": 2070815765.0, "reward": 0.00390625, "reward_std": 0.015625, "rewards/accuracy_reward/mean": 0.00390625, "rewards/accuracy_reward/std": 0.06243881583213806, "step": 2177 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.65625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2015.0, "completions/mean_length": 1584.033203125, "completions/mean_terminated_length": 698.2784423828125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7435350345651617, "frac_reward_zero_std": 0.78125, "grad_norm": 24.690880484134997, "learning_rate": 2.6924999170722743e-07, "loss": 0.0142, "num_tokens": 2071700918.0, "reward": 0.06640625, "reward_std": 0.08917921781539917, "rewards/accuracy_reward/mean": 0.06854838877916336, "rewards/accuracy_reward/std": 0.25293973088264465, "step": 2178 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.62109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1943.0, "completions/mean_length": 1509.478515625, "completions/mean_terminated_length": 627.0360717773438, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.7438764188785525, "frac_reward_zero_std": 0.84375, "grad_norm": 11.72867863873013, "learning_rate": 2.6883121568045197e-07, "loss": 0.0077, "num_tokens": 2072559627.0, "reward": 0.048828125, "reward_std": 0.05782270431518555, "rewards/accuracy_reward/mean": 0.05040322616696358, "rewards/accuracy_reward/std": 0.21899642050266266, "step": 2179 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2035.0, "completions/mean_length": 1509.763671875, "completions/mean_terminated_length": 641.994873046875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7442178031919433, "frac_reward_zero_std": 0.6875, "grad_norm": 9.443728891362426, "learning_rate": 2.6841283872133954e-07, "loss": 0.018, "num_tokens": 2073403794.0, "reward": 0.0390625, "reward_std": 0.09836846590042114, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 2180 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1990.0, "completions/mean_length": 1672.67578125, "completions/mean_terminated_length": 675.7857055664062, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7445591875053341, "frac_reward_zero_std": 0.71875, "grad_norm": 24.265957413288792, "learning_rate": 2.6799486142369955e-07, "loss": 0.0191, "num_tokens": 2074350796.0, "reward": 0.07421875, "reward_std": 0.1152530387043953, "rewards/accuracy_reward/mean": 0.07421875, "rewards/accuracy_reward/std": 0.2623828947544098, "step": 2181 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.626953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1991.0, "completions/mean_length": 1517.095703125, "completions/mean_terminated_length": 624.8429565429688, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.7449005718187249, "frac_reward_zero_std": 0.6875, "grad_norm": 148.11759865013968, "learning_rate": 2.6757728438077414e-07, "loss": 0.0042, "num_tokens": 2075205309.0, "reward": 0.064453125, "reward_std": 0.1170879527926445, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 2182 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1816.0, "completions/mean_length": 1618.65625, "completions/mean_terminated_length": 638.871826171875, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.7452419561321157, "frac_reward_zero_std": 0.78125, "grad_norm": 71.36047709872058, "learning_rate": 2.6716010818523794e-07, "loss": 0.0226, "num_tokens": 2076106925.0, "reward": 0.021484375, "reward_std": 0.06425705552101135, "rewards/accuracy_reward/mean": 0.02217741869390011, "rewards/accuracy_reward/std": 0.14740891754627228, "step": 2183 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.72265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1787.0, "completions/mean_length": 1637.396484375, "completions/mean_terminated_length": 567.5140991210938, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.7455833404455066, "frac_reward_zero_std": 0.6875, "grad_norm": 28.292484266741525, "learning_rate": 2.667433334291958e-07, "loss": 0.0108, "num_tokens": 2077030104.0, "reward": 0.068359375, "reward_std": 0.11910437047481537, "rewards/accuracy_reward/mean": 0.068359375, "rewards/accuracy_reward/std": 0.25260838866233826, "step": 2184 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.673828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1957.0, "completions/mean_length": 1627.892578125, "completions/mean_terminated_length": 760.0060424804688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7459247247588974, "frac_reward_zero_std": 0.75, "grad_norm": 14.0906711081578, "learning_rate": 2.663269607041837e-07, "loss": -0.0025, "num_tokens": 2077936529.0, "reward": 0.041015625, "reward_std": 0.09258585423231125, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 2185 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1877.0, "completions/mean_length": 1548.1015625, "completions/mean_terminated_length": 559.9302368164062, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.7462661090722881, "frac_reward_zero_std": 0.6875, "grad_norm": 55.286040781531426, "learning_rate": 2.6591099060116625e-07, "loss": -0.0048, "num_tokens": 2078810213.0, "reward": 0.048828125, "reward_std": 0.10598184913396835, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 2186 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.67578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2000.0, "completions/mean_length": 1591.287109375, "completions/mean_terminated_length": 639.3433227539062, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7466074933856789, "frac_reward_zero_std": 0.625, "grad_norm": 24.122156358625357, "learning_rate": 2.6549542371053714e-07, "loss": -0.0112, "num_tokens": 2079701944.0, "reward": 0.109375, "reward_std": 0.13574068248271942, "rewards/accuracy_reward/mean": 0.109375, "rewards/accuracy_reward/std": 0.31241437792778015, "step": 2187 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.66015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1925.0, "completions/mean_length": 1615.23828125, "completions/mean_terminated_length": 774.586181640625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7469488776990697, "frac_reward_zero_std": 0.71875, "grad_norm": 40.63971003997408, "learning_rate": 2.650802606221175e-07, "loss": 0.0374, "num_tokens": 2080605314.0, "reward": 0.103515625, "reward_std": 0.11122028529644012, "rewards/accuracy_reward/mean": 0.103515625, "rewards/accuracy_reward/std": 0.30492907762527466, "step": 2188 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.708984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1894.0, "completions/mean_length": 1627.2890625, "completions/mean_terminated_length": 602.3355712890625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7472902620124605, "frac_reward_zero_std": 0.71875, "grad_norm": 13.954175533464152, "learning_rate": 2.6466550192515526e-07, "loss": 0.0136, "num_tokens": 2081513350.0, "reward": 0.080078125, "reward_std": 0.11372203379869461, "rewards/accuracy_reward/mean": 0.080078125, "rewards/accuracy_reward/std": 0.271679550409317, "step": 2189 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.638671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2048.0, "completions/mean_length": 1549.09765625, "completions/mean_terminated_length": 667.2540893554688, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.7476316463258513, "frac_reward_zero_std": 0.75, "grad_norm": 57.169619310039074, "learning_rate": 2.642511482083247e-07, "loss": 0.0219, "num_tokens": 2082380552.0, "reward": 0.064453125, "reward_std": 0.10296767204999924, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 2190 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.697265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2015.0, "completions/mean_length": 1617.09375, "completions/mean_terminated_length": 624.6193237304688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7479730306392421, "frac_reward_zero_std": 0.75, "grad_norm": 31.768027290790332, "learning_rate": 2.638372000597251e-07, "loss": -0.0031, "num_tokens": 2083288728.0, "reward": 0.052734375, "reward_std": 0.0985880047082901, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 2191 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.68359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1943.0, "completions/mean_length": 1601.693359375, "completions/mean_terminated_length": 637.4506225585938, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.748314414952633, "frac_reward_zero_std": 0.75, "grad_norm": 23.65628818225515, "learning_rate": 2.634236580668802e-07, "loss": 0.0163, "num_tokens": 2084188651.0, "reward": 0.05078125, "reward_std": 0.09775100648403168, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 2192 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1883.0, "completions/mean_length": 1674.216796875, "completions/mean_terminated_length": 598.1742553710938, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7486557992660238, "frac_reward_zero_std": 0.90625, "grad_norm": 0.21046703825872332, "learning_rate": 2.630105228167369e-07, "loss": -0.0013, "num_tokens": 2085113130.0, "reward": 0.02734375, "reward_std": 0.02960042469203472, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 2193 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.673828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1800.0, "completions/mean_length": 1591.76953125, "completions/mean_terminated_length": 649.2575073242188, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.7489971835794145, "frac_reward_zero_std": 0.78125, "grad_norm": 23.454985233805438, "learning_rate": 2.625977948956656e-07, "loss": 0.0042, "num_tokens": 2086005556.0, "reward": 0.048828125, "reward_std": 0.07328139245510101, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 2194 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1820.0, "completions/mean_length": 1602.3203125, "completions/mean_terminated_length": 463.75, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.7493385678928053, "frac_reward_zero_std": 0.75, "grad_norm": 35.59483977039, "learning_rate": 2.621854748894578e-07, "loss": -0.0134, "num_tokens": 2086908360.0, "reward": 0.05859375, "reward_std": 0.08985796570777893, "rewards/accuracy_reward/mean": 0.060483869165182114, "rewards/accuracy_reward/std": 0.2386218160390854, "step": 2195 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.73046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1679.0, "completions/mean_length": 1669.181640625, "completions/mean_terminated_length": 642.5289916992188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7496799522061961, "frac_reward_zero_std": 0.8125, "grad_norm": 23.94870590034524, "learning_rate": 2.6177356338332635e-07, "loss": 0.0088, "num_tokens": 2087844149.0, "reward": 0.02734375, "reward_std": 0.06233368441462517, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 2196 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1590.728515625, "completions/mean_terminated_length": 584.7312622070312, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7500213365195869, "frac_reward_zero_std": 0.59375, "grad_norm": 24.11278995839209, "learning_rate": 2.6136206096190445e-07, "loss": 0.0136, "num_tokens": 2088740170.0, "reward": 0.11328125, "reward_std": 0.1556394398212433, "rewards/accuracy_reward/mean": 0.11328125, "rewards/accuracy_reward/std": 0.3172462284564972, "step": 2197 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.697265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1899.0, "completions/mean_length": 1629.341796875, "completions/mean_terminated_length": 665.077392578125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.7503627208329777, "frac_reward_zero_std": 0.6875, "grad_norm": 15.749584835939213, "learning_rate": 2.609509682092442e-07, "loss": 0.0256, "num_tokens": 2089648905.0, "reward": 0.099609375, "reward_std": 0.11646249890327454, "rewards/accuracy_reward/mean": 0.099609375, "rewards/accuracy_reward/std": 0.29977133870124817, "step": 2198 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.689453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2007.0, "completions/mean_length": 1607.853515625, "completions/mean_terminated_length": 630.6729125976562, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7507041051463685, "frac_reward_zero_std": 0.75, "grad_norm": 23.210973572572502, "learning_rate": 2.6054028570881697e-07, "loss": 0.0308, "num_tokens": 2090544270.0, "reward": 0.076171875, "reward_std": 0.1091160923242569, "rewards/accuracy_reward/mean": 0.0786290317773819, "rewards/accuracy_reward/std": 0.26943066716194153, "step": 2199 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.642578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1818.0, "completions/mean_length": 1560.884765625, "completions/mean_terminated_length": 685.1420288085938, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7510454894597594, "frac_reward_zero_std": 0.71875, "grad_norm": 299.73813234765373, "learning_rate": 2.6013001404351133e-07, "loss": 0.0248, "num_tokens": 2091415171.0, "reward": 0.0625, "reward_std": 0.09959383308887482, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.2422981858253479, "step": 2200 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.673828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1858.0, "completions/mean_length": 1600.435546875, "completions/mean_terminated_length": 675.8263549804688, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.7513868737731502, "frac_reward_zero_std": 0.78125, "grad_norm": 9.78397215096451, "learning_rate": 2.5972015379563263e-07, "loss": 0.0217, "num_tokens": 2092311362.0, "reward": 0.041015625, "reward_std": 0.080595001578331, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 2201 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1856.0, "completions/mean_length": 1500.95703125, "completions/mean_terminated_length": 589.21875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7517282580865409, "frac_reward_zero_std": 0.71875, "grad_norm": 52.00350531787857, "learning_rate": 2.5931070554690284e-07, "loss": -0.0079, "num_tokens": 2093148780.0, "reward": 0.048828125, "reward_std": 0.09673243016004562, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 2202 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.681640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1973.0, "completions/mean_length": 1604.19921875, "completions/mean_terminated_length": 653.9754638671875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.7520696423999317, "frac_reward_zero_std": 0.59375, "grad_norm": 30.421510031241386, "learning_rate": 2.589016698784585e-07, "loss": 0.0474, "num_tokens": 2094041970.0, "reward": 0.078125, "reward_std": 0.15327692031860352, "rewards/accuracy_reward/mean": 0.078125, "rewards/accuracy_reward/std": 0.26863065361976624, "step": 2203 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.794921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2045.0, "completions/mean_length": 1790.36328125, "completions/mean_terminated_length": 791.7142944335938, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7524110267133225, "frac_reward_zero_std": 0.78125, "grad_norm": 20.689079597521534, "learning_rate": 2.5849304737085143e-07, "loss": 0.0177, "num_tokens": 2095040220.0, "reward": 0.05859375, "reward_std": 0.08522041887044907, "rewards/accuracy_reward/mean": 0.05859375, "rewards/accuracy_reward/std": 0.23509246110916138, "step": 2204 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.685546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1882.0, "completions/mean_length": 1585.25390625, "completions/mean_terminated_length": 576.4099731445312, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7527524110267133, "frac_reward_zero_std": 0.90625, "grad_norm": 10.800193253984355, "learning_rate": 2.5808483860404605e-07, "loss": 0.0155, "num_tokens": 2095925854.0, "reward": 0.017578125, "reward_std": 0.03630761057138443, "rewards/accuracy_reward/mean": 0.017578125, "rewards/accuracy_reward/std": 0.13154059648513794, "step": 2205 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2024.0, "completions/mean_length": 1612.21875, "completions/mean_terminated_length": 653.5, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.7530937953401041, "frac_reward_zero_std": 0.75, "grad_norm": 76.97829901393658, "learning_rate": 2.576770441574204e-07, "loss": 0.014, "num_tokens": 2096829726.0, "reward": 0.095703125, "reward_std": 0.10100477933883667, "rewards/accuracy_reward/mean": 0.09879032522439957, "rewards/accuracy_reward/std": 0.2986815273761749, "step": 2206 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.69921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1942.0, "completions/mean_length": 1616.7734375, "completions/mean_terminated_length": 614.3117065429688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7534351796534949, "frac_reward_zero_std": 0.65625, "grad_norm": 36.333292473213305, "learning_rate": 2.5726966460976406e-07, "loss": 0.0204, "num_tokens": 2097735418.0, "reward": 0.09375, "reward_std": 0.1331530660390854, "rewards/accuracy_reward/mean": 0.09375, "rewards/accuracy_reward/std": 0.29176566004753113, "step": 2207 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.638671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2001.0, "completions/mean_length": 1518.87890625, "completions/mean_terminated_length": 583.6216430664062, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7537765639668857, "frac_reward_zero_std": 0.84375, "grad_norm": 31.126770946411682, "learning_rate": 2.5686270053927743e-07, "loss": 0.0057, "num_tokens": 2098596396.0, "reward": 0.0234375, "reward_std": 0.05721627548336983, "rewards/accuracy_reward/mean": 0.0234375, "rewards/accuracy_reward/std": 0.15143637359142303, "step": 2208 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1866.0, "completions/mean_length": 1621.298828125, "completions/mean_terminated_length": 571.8446044921875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7541179482802766, "frac_reward_zero_std": 0.78125, "grad_norm": 27.15206895078571, "learning_rate": 2.5645615252357205e-07, "loss": 0.0393, "num_tokens": 2099508501.0, "reward": 0.056640625, "reward_std": 0.08804761618375778, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 2209 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.677734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1961.0, "completions/mean_length": 1586.888671875, "completions/mean_terminated_length": 617.4909057617188, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.7544593325936674, "frac_reward_zero_std": 0.84375, "grad_norm": 29.903027516589866, "learning_rate": 2.560500211396681e-07, "loss": 0.0087, "num_tokens": 2100402124.0, "reward": 0.025390625, "reward_std": 0.05452118441462517, "rewards/accuracy_reward/mean": 0.025390625, "rewards/accuracy_reward/std": 0.15746226906776428, "step": 2210 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.728515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1986.0, "completions/mean_length": 1650.51953125, "completions/mean_terminated_length": 583.8992919921875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7548007169070581, "frac_reward_zero_std": 0.8125, "grad_norm": 17.082855226278348, "learning_rate": 2.556443069639951e-07, "loss": 0.0098, "num_tokens": 2101318790.0, "reward": 0.037109375, "reward_std": 0.07085913419723511, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 2211 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1837.0, "completions/mean_length": 1662.154296875, "completions/mean_terminated_length": 595.4044189453125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7551421012204489, "frac_reward_zero_std": 0.71875, "grad_norm": 34.739905982749036, "learning_rate": 2.5523901057238994e-07, "loss": 0.0122, "num_tokens": 2102249525.0, "reward": 0.060546875, "reward_std": 0.10960254818201065, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 2212 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.65625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1918.0, "completions/mean_length": 1562.990234375, "completions/mean_terminated_length": 637.0625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7554834855338397, "frac_reward_zero_std": 0.8125, "grad_norm": 63.57129223584638, "learning_rate": 2.5483413254009666e-07, "loss": 0.0078, "num_tokens": 2103123360.0, "reward": 0.029296875, "reward_std": 0.06271954625844955, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 2213 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1996.0, "completions/mean_length": 1557.833984375, "completions/mean_terminated_length": 653.75, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7558248698472305, "frac_reward_zero_std": 0.71875, "grad_norm": 20.925509821042002, "learning_rate": 2.544296734417658e-07, "loss": 0.0196, "num_tokens": 2103999483.0, "reward": 0.095703125, "reward_std": 0.11623741686344147, "rewards/accuracy_reward/mean": 0.095703125, "rewards/accuracy_reward/std": 0.2944713830947876, "step": 2214 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.669921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1997.0, "completions/mean_length": 1572.49609375, "completions/mean_terminated_length": 607.4201049804688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7561662541606213, "frac_reward_zero_std": 0.75, "grad_norm": 7.114035088567925, "learning_rate": 2.540256338514528e-07, "loss": -0.0115, "num_tokens": 2104874457.0, "reward": 0.05078125, "reward_std": 0.08746850490570068, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 2215 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.69921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2002.0, "completions/mean_length": 1611.82421875, "completions/mean_terminated_length": 597.8571166992188, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7565076384740121, "frac_reward_zero_std": 0.75, "grad_norm": 26.20568632094594, "learning_rate": 2.536220143426182e-07, "loss": 0.0357, "num_tokens": 2105772319.0, "reward": 0.09375, "reward_std": 0.11630973219871521, "rewards/accuracy_reward/mean": 0.09677419066429138, "rewards/accuracy_reward/std": 0.2959485352039337, "step": 2216 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.701171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2025.0, "completions/mean_length": 1610.703125, "completions/mean_terminated_length": 584.62744140625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.756849022787403, "frac_reward_zero_std": 0.65625, "grad_norm": 42.62162106043469, "learning_rate": 2.532188154881258e-07, "loss": 0.0013, "num_tokens": 2106669223.0, "reward": 0.08203125, "reward_std": 0.12754468619823456, "rewards/accuracy_reward/mean": 0.08203125, "rewards/accuracy_reward/std": 0.2746807038784027, "step": 2217 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.751953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1943.0, "completions/mean_length": 1688.08984375, "completions/mean_terminated_length": 597.0236206054688, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7571904071007938, "frac_reward_zero_std": 0.65625, "grad_norm": 113.5196805740244, "learning_rate": 2.528160378602431e-07, "loss": 0.0208, "num_tokens": 2107605237.0, "reward": 0.068359375, "reward_std": 0.11685723811388016, "rewards/accuracy_reward/mean": 0.068359375, "rewards/accuracy_reward/std": 0.25260838866233826, "step": 2218 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.732421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2010.0, "completions/mean_length": 1667.0546875, "completions/mean_terminated_length": 624.7372436523438, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.7575317914141845, "frac_reward_zero_std": 0.78125, "grad_norm": 85.14756383016665, "learning_rate": 2.5241368203063875e-07, "loss": -0.0011, "num_tokens": 2108532369.0, "reward": 0.046875, "reward_std": 0.08670367300510406, "rewards/accuracy_reward/mean": 0.046875, "rewards/accuracy_reward/std": 0.21157780289649963, "step": 2219 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.73828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1784.0, "completions/mean_length": 1672.328125, "completions/mean_terminated_length": 612.5969848632812, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7578731757275753, "frac_reward_zero_std": 0.65625, "grad_norm": 116.23516740238533, "learning_rate": 2.5201174857038344e-07, "loss": 0.0042, "num_tokens": 2109469753.0, "reward": 0.04296875, "reward_std": 0.11123858392238617, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 2220 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.70703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1829.0, "completions/mean_length": 1617.462890625, "completions/mean_terminated_length": 578.433349609375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7582145600409661, "frac_reward_zero_std": 0.75, "grad_norm": 29.659560851519238, "learning_rate": 2.516102380499483e-07, "loss": -0.0018, "num_tokens": 2110374358.0, "reward": 0.078125, "reward_std": 0.10150270164012909, "rewards/accuracy_reward/mean": 0.078125, "rewards/accuracy_reward/std": 0.26863065361976624, "step": 2221 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.744140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1997.0, "completions/mean_length": 1695.029296875, "completions/mean_terminated_length": 668.4503784179688, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.7585559443543569, "frac_reward_zero_std": 0.71875, "grad_norm": 50.878008347479955, "learning_rate": 2.51209151039204e-07, "loss": 0.0238, "num_tokens": 2111319525.0, "reward": 0.048828125, "reward_std": 0.09143522381782532, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 2222 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.751953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1993.0, "completions/mean_length": 1700.81640625, "completions/mean_terminated_length": 648.3306884765625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7588973286677477, "frac_reward_zero_std": 0.8125, "grad_norm": 90.00534449616639, "learning_rate": 2.5080848810742027e-07, "loss": -0.0093, "num_tokens": 2112276359.0, "reward": 0.0390625, "reward_std": 0.07187044620513916, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 2223 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.736328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1936.0, "completions/mean_length": 1665.08203125, "completions/mean_terminated_length": 595.7481079101562, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7592387129811385, "frac_reward_zero_std": 0.75, "grad_norm": 27.476201648315932, "learning_rate": 2.504082498232648e-07, "loss": 0.0002, "num_tokens": 2113199985.0, "reward": 0.048828125, "reward_std": 0.09072227776050568, "rewards/accuracy_reward/mean": 0.05040322616696358, "rewards/accuracy_reward/std": 0.21899642050266266, "step": 2224 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1989.0, "completions/mean_length": 1631.884765625, "completions/mean_terminated_length": 682.2884521484375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7595800972945294, "frac_reward_zero_std": 0.84375, "grad_norm": 21.959572119266443, "learning_rate": 2.5000843675480264e-07, "loss": -0.0017, "num_tokens": 2114109126.0, "reward": 0.013671875, "reward_std": 0.043847277760505676, "rewards/accuracy_reward/mean": 0.013671875, "rewards/accuracy_reward/std": 0.1162383034825325, "step": 2225 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.69140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1961.0, "completions/mean_length": 1618.904296875, "completions/mean_terminated_length": 657.5126953125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7599214816079202, "frac_reward_zero_std": 0.75, "grad_norm": 9.159964388175874, "learning_rate": 2.4960904946949513e-07, "loss": 0.0114, "num_tokens": 2115011749.0, "reward": 0.0625, "reward_std": 0.103780098259449, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.2422981858253479, "step": 2226 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.693359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2015.0, "completions/mean_length": 1614.96484375, "completions/mean_terminated_length": 635.8088989257812, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7602628659213109, "frac_reward_zero_std": 0.71875, "grad_norm": 19.31017129891979, "learning_rate": 2.492100885341997e-07, "loss": -0.0014, "num_tokens": 2115922643.0, "reward": 0.072265625, "reward_std": 0.09886875748634338, "rewards/accuracy_reward/mean": 0.072265625, "rewards/accuracy_reward/std": 0.2591804563999176, "step": 2227 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.654296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1995.0, "completions/mean_length": 1553.955078125, "completions/mean_terminated_length": 618.8983154296875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7606042502347017, "frac_reward_zero_std": 0.90625, "grad_norm": 16.480227990445243, "learning_rate": 2.4881155451516844e-07, "loss": 0.0135, "num_tokens": 2116791036.0, "reward": 0.021484375, "reward_std": 0.03462383896112442, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 2228 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.712890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1713.0, "completions/mean_length": 1627.5234375, "completions/mean_terminated_length": 583.4829711914062, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.7609456345480925, "frac_reward_zero_std": 0.78125, "grad_norm": 19.099363321025816, "learning_rate": 2.484134479780473e-07, "loss": 0.0081, "num_tokens": 2117700504.0, "reward": 0.03125, "reward_std": 0.07559756934642792, "rewards/accuracy_reward/mean": 0.032258063554763794, "rewards/accuracy_reward/std": 0.17686307430267334, "step": 2229 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1947.0, "completions/mean_length": 1667.388671875, "completions/mean_terminated_length": 571.6893920898438, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.7612870188614833, "frac_reward_zero_std": 0.6875, "grad_norm": 128.03170989470343, "learning_rate": 2.48015769487876e-07, "loss": -0.0015, "num_tokens": 2118630031.0, "reward": 0.056640625, "reward_std": 0.11492595076560974, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 2230 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1815.0, "completions/mean_length": 1628.451171875, "completions/mean_terminated_length": 596.5878295898438, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7616284031748741, "frac_reward_zero_std": 0.84375, "grad_norm": 42.09742917566184, "learning_rate": 2.476185196090862e-07, "loss": 0.0145, "num_tokens": 2119536438.0, "reward": 0.048828125, "reward_std": 0.05688370764255524, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 2231 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.775390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1958.0, "completions/mean_length": 1735.60546875, "completions/mean_terminated_length": 657.6956176757812, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7619697874882649, "frac_reward_zero_std": 0.8125, "grad_norm": 5.257373094538761, "learning_rate": 2.472216989055015e-07, "loss": 0.0237, "num_tokens": 2120497532.0, "reward": 0.037109375, "reward_std": 0.061668481677770615, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 2232 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.755859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1862.0, "completions/mean_length": 1700.455078125, "completions/mean_terminated_length": 624.4560546875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7623111718016558, "frac_reward_zero_std": 0.71875, "grad_norm": 89.81193074808552, "learning_rate": 2.468253079403362e-07, "loss": 0.0062, "num_tokens": 2121458517.0, "reward": 0.060546875, "reward_std": 0.10271549224853516, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 2233 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1999.0, "completions/mean_length": 1619.662109375, "completions/mean_terminated_length": 642.173095703125, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.7626525561150466, "frac_reward_zero_std": 0.75, "grad_norm": 16.615257359413665, "learning_rate": 2.464293472761948e-07, "loss": -0.0042, "num_tokens": 2122369640.0, "reward": 0.037109375, "reward_std": 0.08537977933883667, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 2234 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.70703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1542.0, "completions/mean_length": 1611.814453125, "completions/mean_terminated_length": 559.1533203125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7629939404284373, "frac_reward_zero_std": 0.65625, "grad_norm": 12.452921701009723, "learning_rate": 2.460338174750713e-07, "loss": 0.0235, "num_tokens": 2123273241.0, "reward": 0.076171875, "reward_std": 0.1247420534491539, "rewards/accuracy_reward/mean": 0.0786290317773819, "rewards/accuracy_reward/std": 0.26943066716194153, "step": 2235 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.708984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1860.0, "completions/mean_length": 1633.9765625, "completions/mean_terminated_length": 625.6778564453125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7633353247418281, "frac_reward_zero_std": 0.75, "grad_norm": 61.45202360904577, "learning_rate": 2.4563871909834755e-07, "loss": 0.0027, "num_tokens": 2124195325.0, "reward": 0.052734375, "reward_std": 0.0881948471069336, "rewards/accuracy_reward/mean": 0.05443548411130905, "rewards/accuracy_reward/std": 0.227104052901268, "step": 2236 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.669921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1825.0, "completions/mean_length": 1561.017578125, "completions/mean_terminated_length": 572.6449584960938, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7636767090552189, "frac_reward_zero_std": 0.65625, "grad_norm": 29.91338125473617, "learning_rate": 2.4524405270679386e-07, "loss": -0.0119, "num_tokens": 2125075142.0, "reward": 0.091796875, "reward_std": 0.12402814626693726, "rewards/accuracy_reward/mean": 0.091796875, "rewards/accuracy_reward/std": 0.289021372795105, "step": 2237 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.689453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2013.0, "completions/mean_length": 1587.421875, "completions/mean_terminated_length": 564.8804931640625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.7640180933686097, "frac_reward_zero_std": 0.75, "grad_norm": 54.1485772617581, "learning_rate": 2.4484981886056647e-07, "loss": 0.0221, "num_tokens": 2125970958.0, "reward": 0.09375, "reward_std": 0.11018751561641693, "rewards/accuracy_reward/mean": 0.09375, "rewards/accuracy_reward/std": 0.29176566004753113, "step": 2238 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.681640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1799.0, "completions/mean_length": 1620.25390625, "completions/mean_terminated_length": 704.4049072265625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7643594776820005, "frac_reward_zero_std": 0.78125, "grad_norm": 93.1376682315806, "learning_rate": 2.444560181192087e-07, "loss": 0.0118, "num_tokens": 2126874496.0, "reward": 0.052734375, "reward_std": 0.08505964279174805, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 2239 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.662109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2015.0, "completions/mean_length": 1562.755859375, "completions/mean_terminated_length": 611.9017333984375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7647008619953913, "frac_reward_zero_std": 0.71875, "grad_norm": 24.055875642354074, "learning_rate": 2.4406265104164814e-07, "loss": -0.0064, "num_tokens": 2127754371.0, "reward": 0.060546875, "reward_std": 0.10672760009765625, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 2240 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.751953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1968.0, "completions/mean_length": 1684.4375, "completions/mean_terminated_length": 582.2991943359375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7650422463087821, "frac_reward_zero_std": 0.8125, "grad_norm": 71.7981845041056, "learning_rate": 2.4366971818619785e-07, "loss": 0.0048, "num_tokens": 2128689859.0, "reward": 0.033203125, "reward_std": 0.06975477933883667, "rewards/accuracy_reward/mean": 0.033203125, "rewards/accuracy_reward/std": 0.17934183776378632, "step": 2241 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.75390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1929.0, "completions/mean_length": 1699.46875, "completions/mean_terminated_length": 631.74609375, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.765383630622173, "frac_reward_zero_std": 0.625, "grad_norm": 24.872947986499543, "learning_rate": 2.4327722011055407e-07, "loss": 0.0042, "num_tokens": 2129642835.0, "reward": 0.087890625, "reward_std": 0.1310567855834961, "rewards/accuracy_reward/mean": 0.087890625, "rewards/accuracy_reward/std": 0.2834126651287079, "step": 2242 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.72265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1887.0, "completions/mean_length": 1641.755859375, "completions/mean_terminated_length": 583.2323608398438, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7657250149355637, "frac_reward_zero_std": 0.78125, "grad_norm": 15.046160899195035, "learning_rate": 2.428851573717961e-07, "loss": 0.0147, "num_tokens": 2130561974.0, "reward": 0.01953125, "reward_std": 0.06085042655467987, "rewards/accuracy_reward/mean": 0.01953125, "rewards/accuracy_reward/std": 0.1385180652141571, "step": 2243 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2043.0, "completions/mean_length": 1763.599609375, "completions/mean_terminated_length": 700.1851806640625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7660663992489545, "frac_reward_zero_std": 0.6875, "grad_norm": 41.293665093029254, "learning_rate": 2.42493530526385e-07, "loss": 0.0042, "num_tokens": 2131550185.0, "reward": 0.046875, "reward_std": 0.1049705296754837, "rewards/accuracy_reward/mean": 0.046875, "rewards/accuracy_reward/std": 0.21157780289649963, "step": 2244 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.666015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1940.0, "completions/mean_length": 1577.033203125, "completions/mean_terminated_length": 637.8538208007812, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7664077835623453, "frac_reward_zero_std": 0.78125, "grad_norm": 11.66245887058529, "learning_rate": 2.421023401301636e-07, "loss": 0.0052, "num_tokens": 2132438378.0, "reward": 0.025390625, "reward_std": 0.06755761057138443, "rewards/accuracy_reward/mean": 0.026209676638245583, "rewards/accuracy_reward/std": 0.1599196344614029, "step": 2245 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.740234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2030.0, "completions/mean_length": 1678.810546875, "completions/mean_terminated_length": 627.481201171875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7667491678757361, "frac_reward_zero_std": 0.65625, "grad_norm": 41.12605647594873, "learning_rate": 2.417115867383553e-07, "loss": 0.0109, "num_tokens": 2133386265.0, "reward": 0.060546875, "reward_std": 0.12254488468170166, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 2246 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.69140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1909.0, "completions/mean_length": 1580.08984375, "completions/mean_terminated_length": 531.7341918945312, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7670905521891269, "frac_reward_zero_std": 0.8125, "grad_norm": 17.17961463986252, "learning_rate": 2.4132127090556265e-07, "loss": 0.0057, "num_tokens": 2134272871.0, "reward": 0.046875, "reward_std": 0.0687694400548935, "rewards/accuracy_reward/mean": 0.046875, "rewards/accuracy_reward/std": 0.21157780289649963, "step": 2247 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.751953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1881.0, "completions/mean_length": 1694.26171875, "completions/mean_terminated_length": 622.3228149414062, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.7674319365025177, "frac_reward_zero_std": 0.78125, "grad_norm": 62.936927524265975, "learning_rate": 2.4093139318576793e-07, "loss": 0.0311, "num_tokens": 2135221981.0, "reward": 0.078125, "reward_std": 0.08098085969686508, "rewards/accuracy_reward/mean": 0.078125, "rewards/accuracy_reward/std": 0.26863065361976624, "step": 2248 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1898.0, "completions/mean_length": 1594.37109375, "completions/mean_terminated_length": 559.1666870117188, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7677733208159085, "frac_reward_zero_std": 0.6875, "grad_norm": 101.35861448461155, "learning_rate": 2.405419541323314e-07, "loss": 0.0205, "num_tokens": 2136115739.0, "reward": 0.05078125, "reward_std": 0.11585800349712372, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 2249 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.708984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2031.0, "completions/mean_length": 1642.927734375, "completions/mean_terminated_length": 656.4563598632812, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7681147051292994, "frac_reward_zero_std": 0.75, "grad_norm": 49.48775176084035, "learning_rate": 2.4015295429799e-07, "loss": 0.0095, "num_tokens": 2137032454.0, "reward": 0.03125, "reward_std": 0.07795868813991547, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 2250 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.716796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1996.0, "completions/mean_length": 1652.755859375, "completions/mean_terminated_length": 652.3793334960938, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7684560894426901, "frac_reward_zero_std": 0.8125, "grad_norm": 10.934399068283161, "learning_rate": 2.397643942348584e-07, "loss": 0.0237, "num_tokens": 2137955817.0, "reward": 0.0390625, "reward_std": 0.06068410724401474, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 2251 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1639.0, "completions/mean_length": 1675.205078125, "completions/mean_terminated_length": 602.007568359375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7687974737560809, "frac_reward_zero_std": 0.84375, "grad_norm": 8.562930842663347, "learning_rate": 2.393762744944261e-07, "loss": 0.0029, "num_tokens": 2138893970.0, "reward": 0.021484375, "reward_std": 0.05341683328151703, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 2252 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.740234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1996.0, "completions/mean_length": 1689.5, "completions/mean_terminated_length": 667.9097900390625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.7691388580694717, "frac_reward_zero_std": 0.75, "grad_norm": 30.13805288560358, "learning_rate": 2.389885956275585e-07, "loss": 0.0108, "num_tokens": 2139847634.0, "reward": 0.037109375, "reward_std": 0.08780203759670258, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 2253 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2026.0, "completions/mean_length": 1621.458984375, "completions/mean_terminated_length": 683.0687866210938, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7694802423828625, "frac_reward_zero_std": 0.8125, "grad_norm": 38.450479179537155, "learning_rate": 2.386013581844945e-07, "loss": 0.0146, "num_tokens": 2140755613.0, "reward": 0.04296875, "reward_std": 0.06469620764255524, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 2254 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1916.0, "completions/mean_length": 1580.37890625, "completions/mean_terminated_length": 551.6124877929688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7698216266962533, "frac_reward_zero_std": 0.71875, "grad_norm": 6.423560684812626, "learning_rate": 2.3821456271484704e-07, "loss": 0.0256, "num_tokens": 2141633007.0, "reward": 0.04296875, "reward_std": 0.09486784040927887, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 2255 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.681640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1816.0, "completions/mean_length": 1614.126953125, "completions/mean_terminated_length": 685.1594848632812, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7701630110096441, "frac_reward_zero_std": 0.75, "grad_norm": 31.229647576363117, "learning_rate": 2.3782820976760153e-07, "loss": 0.0115, "num_tokens": 2142540320.0, "reward": 0.05078125, "reward_std": 0.09671792387962341, "rewards/accuracy_reward/mean": 0.052419353276491165, "rewards/accuracy_reward/std": 0.22309619188308716, "step": 2256 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.794921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1864.0, "completions/mean_length": 1735.603515625, "completions/mean_terminated_length": 524.6952514648438, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7705043953230349, "frac_reward_zero_std": 0.78125, "grad_norm": 94.13466649496164, "learning_rate": 2.374422998911151e-07, "loss": -0.0047, "num_tokens": 2143504469.0, "reward": 0.033203125, "reward_std": 0.07586899399757385, "rewards/accuracy_reward/mean": 0.033203125, "rewards/accuracy_reward/std": 0.17934183776378632, "step": 2257 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.708984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2020.0, "completions/mean_length": 1626.51953125, "completions/mean_terminated_length": 599.6912841796875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7708457796364258, "frac_reward_zero_std": 0.78125, "grad_norm": 26.25394798125014, "learning_rate": 2.3705683363311656e-07, "loss": 0.0085, "num_tokens": 2144411375.0, "reward": 0.048828125, "reward_std": 0.06536141037940979, "rewards/accuracy_reward/mean": 0.05040322616696358, "rewards/accuracy_reward/std": 0.21899642050266266, "step": 2258 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.65625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1992.0, "completions/mean_length": 1566.763671875, "completions/mean_terminated_length": 648.039794921875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7711871639498165, "frac_reward_zero_std": 0.6875, "grad_norm": 116.8396280220696, "learning_rate": 2.3667181154070443e-07, "loss": 0.0289, "num_tokens": 2145281910.0, "reward": 0.056640625, "reward_std": 0.11036626249551773, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 2259 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.77734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1947.0, "completions/mean_length": 1728.2890625, "completions/mean_terminated_length": 612.1052856445312, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7715285482632073, "frac_reward_zero_std": 0.6875, "grad_norm": 13.737625679621692, "learning_rate": 2.3628723416034742e-07, "loss": 0.0133, "num_tokens": 2146241626.0, "reward": 0.064453125, "reward_std": 0.11982375383377075, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 2260 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.751953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1962.0, "completions/mean_length": 1688.43359375, "completions/mean_terminated_length": 598.409423828125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7718699325765981, "frac_reward_zero_std": 0.71875, "grad_norm": 7.1005908054538525, "learning_rate": 2.359031020378827e-07, "loss": 0.0258, "num_tokens": 2147174296.0, "reward": 0.04296875, "reward_std": 0.09392979741096497, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 2261 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.69921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2002.0, "completions/mean_length": 1648.703125, "completions/mean_terminated_length": 720.467529296875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7722113168899889, "frac_reward_zero_std": 0.78125, "grad_norm": 45.68550030411381, "learning_rate": 2.3551941571851534e-07, "loss": 0.0112, "num_tokens": 2148090000.0, "reward": 0.025390625, "reward_std": 0.06943464279174805, "rewards/accuracy_reward/mean": 0.026209676638245583, "rewards/accuracy_reward/std": 0.1599196493625641, "step": 2262 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.763671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1873.0, "completions/mean_length": 1718.291015625, "completions/mean_terminated_length": 653.322265625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.7725527012033797, "frac_reward_zero_std": 0.8125, "grad_norm": 38.33644412997567, "learning_rate": 2.3513617574681828e-07, "loss": 0.0121, "num_tokens": 2149041413.0, "reward": 0.015625, "reward_std": 0.051659777760505676, "rewards/accuracy_reward/mean": 0.015625, "rewards/accuracy_reward/std": 0.12414088100194931, "step": 2263 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.701171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1978.0, "completions/mean_length": 1625.869140625, "completions/mean_terminated_length": 635.3790893554688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7728940855167705, "frac_reward_zero_std": 0.875, "grad_norm": 18.03017966919862, "learning_rate": 2.347533826667302e-07, "loss": 0.0077, "num_tokens": 2149947922.0, "reward": 0.037109375, "reward_std": 0.046437256038188934, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 2264 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.73046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1754.0, "completions/mean_length": 1663.490234375, "completions/mean_terminated_length": 621.4130249023438, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.7732354698301613, "frac_reward_zero_std": 0.6875, "grad_norm": 20.149293631711878, "learning_rate": 2.3437103702155617e-07, "loss": 0.0083, "num_tokens": 2150869933.0, "reward": 0.10546875, "reward_std": 0.11780644953250885, "rewards/accuracy_reward/mean": 0.10546875, "rewards/accuracy_reward/std": 0.3074568510055542, "step": 2265 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.693359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1889.0, "completions/mean_length": 1570.021484375, "completions/mean_terminated_length": 489.2420349121094, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7735768541435522, "frac_reward_zero_std": 0.75, "grad_norm": 48.9125508012108, "learning_rate": 2.339891393539656e-07, "loss": 0.0292, "num_tokens": 2151746504.0, "reward": 0.083984375, "reward_std": 0.10188952088356018, "rewards/accuracy_reward/mean": 0.083984375, "rewards/accuracy_reward/std": 0.2776356339454651, "step": 2266 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.732421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2048.0, "completions/mean_length": 1686.423828125, "completions/mean_terminated_length": 696.7080078125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7739182384569429, "frac_reward_zero_std": 0.78125, "grad_norm": 22.127961104248037, "learning_rate": 2.336076902059927e-07, "loss": 0.0101, "num_tokens": 2152687441.0, "reward": 0.0546875, "reward_std": 0.08812816441059113, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 2267 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.779296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1917.0, "completions/mean_length": 1727.140625, "completions/mean_terminated_length": 594.1947021484375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7742596227703337, "frac_reward_zero_std": 0.71875, "grad_norm": 76.19844901662259, "learning_rate": 2.3322669011903461e-07, "loss": 0.0021, "num_tokens": 2153647369.0, "reward": 0.052734375, "reward_std": 0.0971580371260643, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 2268 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.775390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1974.0, "completions/mean_length": 1732.42578125, "completions/mean_terminated_length": 643.0086669921875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7746010070837245, "frac_reward_zero_std": 0.71875, "grad_norm": 43.14853579607224, "learning_rate": 2.3284613963385113e-07, "loss": 0.0062, "num_tokens": 2154610899.0, "reward": 0.0390625, "reward_std": 0.09413031488656998, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 2269 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1782.0, "completions/mean_length": 1716.205078125, "completions/mean_terminated_length": 632.3416748046875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7749423913971153, "frac_reward_zero_std": 0.71875, "grad_norm": 68.43877699446037, "learning_rate": 2.3246603929056435e-07, "loss": 0.0086, "num_tokens": 2155563580.0, "reward": 0.0859375, "reward_std": 0.11533089727163315, "rewards/accuracy_reward/mean": 0.0859375, "rewards/accuracy_reward/std": 0.28054583072662354, "step": 2270 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2032.0, "completions/mean_length": 1644.640625, "completions/mean_terminated_length": 572.8571166992188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7752837757105061, "frac_reward_zero_std": 0.90625, "grad_norm": 12.858585896342689, "learning_rate": 2.320863896286569e-07, "loss": -0.0009, "num_tokens": 2156481476.0, "reward": 0.03125, "reward_std": 0.031083684414625168, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 2271 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.689453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1823.0, "completions/mean_length": 1619.005859375, "completions/mean_terminated_length": 666.5848999023438, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7756251600238969, "frac_reward_zero_std": 0.78125, "grad_norm": 44.803110703627276, "learning_rate": 2.3170719118697228e-07, "loss": 0.0069, "num_tokens": 2157384711.0, "reward": 0.048828125, "reward_std": 0.0815330371260643, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 2272 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.732421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1614.0, "completions/mean_length": 1659.51171875, "completions/mean_terminated_length": 596.1314086914062, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7759665443372877, "frac_reward_zero_std": 0.75, "grad_norm": 59.62908370981753, "learning_rate": 2.3132844450371314e-07, "loss": 0.0135, "num_tokens": 2158306685.0, "reward": 0.052734375, "reward_std": 0.09585316479206085, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 2273 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.69921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1978.0, "completions/mean_length": 1629.611328125, "completions/mean_terminated_length": 656.9935302734375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7763079286506785, "frac_reward_zero_std": 0.84375, "grad_norm": 12.710164054319664, "learning_rate": 2.3095015011644128e-07, "loss": 0.0198, "num_tokens": 2159220406.0, "reward": 0.01953125, "reward_std": 0.05243149772286415, "rewards/accuracy_reward/mean": 0.01953125, "rewards/accuracy_reward/std": 0.1385180652141571, "step": 2274 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2030.0, "completions/mean_length": 1696.1484375, "completions/mean_terminated_length": 814.1095581054688, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.7766493129640692, "frac_reward_zero_std": 0.875, "grad_norm": 16.162100384017894, "learning_rate": 2.3057230856207633e-07, "loss": 0.0133, "num_tokens": 2160163202.0, "reward": 0.03125, "reward_std": 0.053257472813129425, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 2275 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1984.0, "completions/mean_length": 1649.50390625, "completions/mean_terminated_length": 650.5342407226562, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7769906972774601, "frac_reward_zero_std": 0.8125, "grad_norm": 69.65292993387105, "learning_rate": 2.3019492037689518e-07, "loss": 0.0034, "num_tokens": 2161081620.0, "reward": 0.0859375, "reward_std": 0.07642117142677307, "rewards/accuracy_reward/mean": 0.0859375, "rewards/accuracy_reward/std": 0.28054583072662354, "step": 2276 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2029.0, "completions/mean_length": 1604.58203125, "completions/mean_terminated_length": 514.0135498046875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7773320815908509, "frac_reward_zero_std": 0.78125, "grad_norm": 13.4489645987364, "learning_rate": 2.2981798609653148e-07, "loss": 0.0078, "num_tokens": 2161973998.0, "reward": 0.0625, "reward_std": 0.09060370922088623, "rewards/accuracy_reward/mean": 0.06451612710952759, "rewards/accuracy_reward/std": 0.2459181249141693, "step": 2277 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.748046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2024.0, "completions/mean_length": 1691.734375, "completions/mean_terminated_length": 633.9844970703125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7776734659042417, "frac_reward_zero_std": 0.90625, "grad_norm": 7.901140110434337, "learning_rate": 2.294415062559743e-07, "loss": -0.0011, "num_tokens": 2162918230.0, "reward": 0.015625, "reward_std": 0.03344620764255524, "rewards/accuracy_reward/mean": 0.015625, "rewards/accuracy_reward/std": 0.12414088100194931, "step": 2278 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.689453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2015.0, "completions/mean_length": 1624.31640625, "completions/mean_terminated_length": 683.6854858398438, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7780148502176325, "frac_reward_zero_std": 0.8125, "grad_norm": 55.95168223664823, "learning_rate": 2.2906548138956815e-07, "loss": -0.0199, "num_tokens": 2163827736.0, "reward": 0.052734375, "reward_std": 0.06260748207569122, "rewards/accuracy_reward/mean": 0.05624999850988388, "rewards/accuracy_reward/std": 0.23064424097537994, "step": 2279 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.708984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2037.0, "completions/mean_length": 1591.580078125, "completions/mean_terminated_length": 479.6308898925781, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7783562345310233, "frac_reward_zero_std": 0.75, "grad_norm": 64.12161728163721, "learning_rate": 2.2868991203101145e-07, "loss": -0.0078, "num_tokens": 2164723505.0, "reward": 0.060546875, "reward_std": 0.10123127698898315, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 2280 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.662109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2044.0, "completions/mean_length": 1587.568359375, "completions/mean_terminated_length": 685.3352661132812, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.7786976188444141, "frac_reward_zero_std": 0.84375, "grad_norm": 184.79244638455157, "learning_rate": 2.283147987133561e-07, "loss": 0.0092, "num_tokens": 2165608084.0, "reward": 0.048828125, "reward_std": 0.06661957502365112, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 2281 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.708984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1792.0, "completions/mean_length": 1607.267578125, "completions/mean_terminated_length": 533.5369262695312, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7790390031578049, "frac_reward_zero_std": 0.78125, "grad_norm": 21.309914748058596, "learning_rate": 2.2794014196900704e-07, "loss": 0.0192, "num_tokens": 2166513277.0, "reward": 0.072265625, "reward_std": 0.08285094052553177, "rewards/accuracy_reward/mean": 0.072265625, "rewards/accuracy_reward/std": 0.2591804563999176, "step": 2282 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.736328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1938.0, "completions/mean_length": 1688.755859375, "completions/mean_terminated_length": 685.5333251953125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7793803874711956, "frac_reward_zero_std": 0.78125, "grad_norm": 25.328743280804456, "learning_rate": 2.275659423297208e-07, "loss": 0.0113, "num_tokens": 2167456480.0, "reward": 0.041015625, "reward_std": 0.07085912674665451, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 2283 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.759765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1768.0, "completions/mean_length": 1703.51171875, "completions/mean_terminated_length": 614.032470703125, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.7797217717845865, "frac_reward_zero_std": 0.8125, "grad_norm": 8.291370927237596, "learning_rate": 2.2719220032660553e-07, "loss": 0.0052, "num_tokens": 2168403238.0, "reward": 0.029296875, "reward_std": 0.06271954625844955, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 2284 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.681640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1755.0, "completions/mean_length": 1580.572265625, "completions/mean_terminated_length": 579.7607421875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7800631560979773, "frac_reward_zero_std": 0.78125, "grad_norm": 12.515069183193397, "learning_rate": 2.2681891649011942e-07, "loss": -0.0059, "num_tokens": 2169291307.0, "reward": 0.05859375, "reward_std": 0.08109388500452042, "rewards/accuracy_reward/mean": 0.05859375, "rewards/accuracy_reward/std": 0.23509246110916138, "step": 2285 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.720703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2009.0, "completions/mean_length": 1670.35546875, "completions/mean_terminated_length": 695.8741455078125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7804045404113681, "frac_reward_zero_std": 0.84375, "grad_norm": 24.44283840250339, "learning_rate": 2.2644609135007088e-07, "loss": 0.0003, "num_tokens": 2170220593.0, "reward": 0.009765625, "reward_std": 0.0390625, "rewards/accuracy_reward/mean": 0.009765625, "rewards/accuracy_reward/std": 0.09843364357948303, "step": 2286 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.69921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1858.0, "completions/mean_length": 1630.17578125, "completions/mean_terminated_length": 658.8701171875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.7807459247247589, "frac_reward_zero_std": 0.78125, "grad_norm": 26.817337426239106, "learning_rate": 2.2607372543561681e-07, "loss": 0.005, "num_tokens": 2171129419.0, "reward": 0.0234375, "reward_std": 0.06657323241233826, "rewards/accuracy_reward/mean": 0.0234375, "rewards/accuracy_reward/std": 0.15143637359142303, "step": 2287 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.755859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1711.0, "completions/mean_length": 1677.357421875, "completions/mean_terminated_length": 529.8480224609375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7810873090381497, "frac_reward_zero_std": 0.75, "grad_norm": 7.395331565369919, "learning_rate": 2.257018192752625e-07, "loss": 0.02, "num_tokens": 2172061730.0, "reward": 0.041015625, "reward_std": 0.08329563587903976, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 2288 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.697265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1874.0, "completions/mean_length": 1613.021484375, "completions/mean_terminated_length": 611.167724609375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7814286933515405, "frac_reward_zero_std": 0.84375, "grad_norm": 23.689557717830898, "learning_rate": 2.2533037339686085e-07, "loss": 0.0187, "num_tokens": 2172971261.0, "reward": 0.0390625, "reward_std": 0.05864076316356659, "rewards/accuracy_reward/mean": 0.04032257944345474, "rewards/accuracy_reward/std": 0.19691328704357147, "step": 2289 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.75, "completions/max_length": 2048.0, "completions/max_terminated_length": 1857.0, "completions/mean_length": 1703.431640625, "completions/mean_terminated_length": 669.7265625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7817700776649313, "frac_reward_zero_std": 0.84375, "grad_norm": 36.07385360632374, "learning_rate": 2.2495938832761114e-07, "loss": -0.0062, "num_tokens": 2173924714.0, "reward": 0.046875, "reward_std": 0.06574726849794388, "rewards/accuracy_reward/mean": 0.046875, "rewards/accuracy_reward/std": 0.21157780289649963, "step": 2290 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.654296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1962.0, "completions/mean_length": 1542.609375, "completions/mean_terminated_length": 586.0791015625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.782111461978322, "frac_reward_zero_std": 0.6875, "grad_norm": 67.19638597512645, "learning_rate": 2.245888645940591e-07, "loss": -0.0012, "num_tokens": 2174784658.0, "reward": 0.103515625, "reward_std": 0.13837699592113495, "rewards/accuracy_reward/mean": 0.103515625, "rewards/accuracy_reward/std": 0.30492907762527466, "step": 2291 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.798828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1886.0, "completions/mean_length": 1761.84765625, "completions/mean_terminated_length": 625.5728149414062, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7824528462917129, "frac_reward_zero_std": 0.875, "grad_norm": 16.926319847930557, "learning_rate": 2.2421880272209524e-07, "loss": 0.0024, "num_tokens": 2175768132.0, "reward": 0.017578125, "reward_std": 0.04461899772286415, "rewards/accuracy_reward/mean": 0.017578125, "rewards/accuracy_reward/std": 0.13154059648513794, "step": 2292 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.736328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2023.0, "completions/mean_length": 1682.759765625, "completions/mean_terminated_length": 662.7926025390625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7827942306051037, "frac_reward_zero_std": 0.75, "grad_norm": 137.8500087942121, "learning_rate": 2.238492032369546e-07, "loss": 0.0058, "num_tokens": 2176710217.0, "reward": 0.07421875, "reward_std": 0.10890313982963562, "rewards/accuracy_reward/mean": 0.07421875, "rewards/accuracy_reward/std": 0.2623828947544098, "step": 2293 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.677734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1998.0, "completions/mean_length": 1578.267578125, "completions/mean_terminated_length": 590.4060668945312, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7831356149184945, "frac_reward_zero_std": 0.75, "grad_norm": 29.109965465128127, "learning_rate": 2.2348006666321633e-07, "loss": 0.0247, "num_tokens": 2177587650.0, "reward": 0.05859375, "reward_std": 0.10270209610462189, "rewards/accuracy_reward/mean": 0.060483869165182114, "rewards/accuracy_reward/std": 0.2386218160390854, "step": 2294 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.708984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1851.0, "completions/mean_length": 1632.9375, "completions/mean_terminated_length": 621.7449951171875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.7834769992318853, "frac_reward_zero_std": 0.875, "grad_norm": 23.843420872438283, "learning_rate": 2.2311139352480196e-07, "loss": 0.0128, "num_tokens": 2178500722.0, "reward": 0.01953125, "reward_std": 0.04318207502365112, "rewards/accuracy_reward/mean": 0.01953125, "rewards/accuracy_reward/std": 0.1385180652141571, "step": 2295 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.74609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2019.0, "completions/mean_length": 1679.150390625, "completions/mean_terminated_length": 595.2999877929688, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7838183835452761, "frac_reward_zero_std": 0.71875, "grad_norm": 51.081935887958686, "learning_rate": 2.227431843449759e-07, "loss": 0.003, "num_tokens": 2179438847.0, "reward": 0.091796875, "reward_std": 0.11261671781539917, "rewards/accuracy_reward/mean": 0.091796875, "rewards/accuracy_reward/std": 0.289021372795105, "step": 2296 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.744140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1859.0, "completions/mean_length": 1701.13671875, "completions/mean_terminated_length": 692.3206176757812, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7841597678586669, "frac_reward_zero_std": 0.78125, "grad_norm": 57.558526105451904, "learning_rate": 2.2237543964634343e-07, "loss": 0.0048, "num_tokens": 2180388837.0, "reward": 0.0859375, "reward_std": 0.09561356902122498, "rewards/accuracy_reward/mean": 0.0859375, "rewards/accuracy_reward/std": 0.28054583072662354, "step": 2297 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.73046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2022.0, "completions/mean_length": 1675.9453125, "completions/mean_terminated_length": 667.6232299804688, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7845011521720577, "frac_reward_zero_std": 0.71875, "grad_norm": 92.13754857222388, "learning_rate": 2.2200815995085132e-07, "loss": 0.0096, "num_tokens": 2181333417.0, "reward": 0.064453125, "reward_std": 0.11182428896427155, "rewards/accuracy_reward/mean": 0.06653226166963577, "rewards/accuracy_reward/std": 0.2494617998600006, "step": 2298 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.755859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2032.0, "completions/mean_length": 1709.51171875, "completions/mean_terminated_length": 661.552001953125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7848425364854484, "frac_reward_zero_std": 0.6875, "grad_norm": 134.59491057686972, "learning_rate": 2.2164134577978528e-07, "loss": 0.0214, "num_tokens": 2182285583.0, "reward": 0.080078125, "reward_std": 0.1344509869813919, "rewards/accuracy_reward/mean": 0.080078125, "rewards/accuracy_reward/std": 0.271679550409317, "step": 2299 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.75390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2017.0, "completions/mean_length": 1689.658203125, "completions/mean_terminated_length": 591.8809814453125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.7851839207988393, "frac_reward_zero_std": 0.84375, "grad_norm": 41.889077917804855, "learning_rate": 2.2127499765377133e-07, "loss": 0.0004, "num_tokens": 2183230864.0, "reward": 0.0234375, "reward_std": 0.05012226849794388, "rewards/accuracy_reward/mean": 0.0234375, "rewards/accuracy_reward/std": 0.15143637359142303, "step": 2300 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.775390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2025.0, "completions/mean_length": 1727.33203125, "completions/mean_terminated_length": 620.3303833007812, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.7855253051122301, "frac_reward_zero_std": 0.84375, "grad_norm": 3.948665500334967, "learning_rate": 2.2090911609277375e-07, "loss": 0.014, "num_tokens": 2184196538.0, "reward": 0.01953125, "reward_std": 0.05001020431518555, "rewards/accuracy_reward/mean": 0.01953125, "rewards/accuracy_reward/std": 0.1385180652141571, "step": 2301 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.728515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1996.0, "completions/mean_length": 1665.33203125, "completions/mean_terminated_length": 638.798583984375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7858666894256209, "frac_reward_zero_std": 0.65625, "grad_norm": 44.789119092323645, "learning_rate": 2.2054370161609415e-07, "loss": 0.0061, "num_tokens": 2185125012.0, "reward": 0.11328125, "reward_std": 0.14328739047050476, "rewards/accuracy_reward/mean": 0.11328125, "rewards/accuracy_reward/std": 0.3172462284564972, "step": 2302 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.720703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1978.0, "completions/mean_length": 1684.41796875, "completions/mean_terminated_length": 746.2237548828125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7862080737390117, "frac_reward_zero_std": 0.71875, "grad_norm": 23.803980615278903, "learning_rate": 2.201787547423719e-07, "loss": 0.0157, "num_tokens": 2186064394.0, "reward": 0.056640625, "reward_std": 0.10541516542434692, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 2303 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.701171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1820.0, "completions/mean_length": 1607.771484375, "completions/mean_terminated_length": 574.8170166015625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7865494580524025, "frac_reward_zero_std": 0.8125, "grad_norm": 10.543524482588994, "learning_rate": 2.19814275989582e-07, "loss": 0.0174, "num_tokens": 2186962181.0, "reward": 0.0703125, "reward_std": 0.08177263289690018, "rewards/accuracy_reward/mean": 0.0703125, "rewards/accuracy_reward/std": 0.25592297315597534, "step": 2304 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.740234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1729.0, "completions/mean_length": 1666.556640625, "completions/mean_terminated_length": 579.5864868164062, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7868908423657933, "frac_reward_zero_std": 0.6875, "grad_norm": 33.94891767565079, "learning_rate": 2.1945026587503578e-07, "loss": 0.0481, "num_tokens": 2187889634.0, "reward": 0.068359375, "reward_std": 0.12241136282682419, "rewards/accuracy_reward/mean": 0.07056451588869095, "rewards/accuracy_reward/std": 0.25635457038879395, "step": 2305 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.67578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1684.0, "completions/mean_length": 1557.46875, "completions/mean_terminated_length": 535.0361328125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7872322266791841, "frac_reward_zero_std": 0.8125, "grad_norm": 18.71055783535664, "learning_rate": 2.1908672491537854e-07, "loss": 0.0085, "num_tokens": 2188764194.0, "reward": 0.046875, "reward_std": 0.06574726849794388, "rewards/accuracy_reward/mean": 0.046875, "rewards/accuracy_reward/std": 0.21157780289649963, "step": 2306 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1957.0, "completions/mean_length": 1718.74609375, "completions/mean_terminated_length": 594.7413940429688, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.787573610992575, "frac_reward_zero_std": 0.625, "grad_norm": 88.38607462571197, "learning_rate": 2.187236536265904e-07, "loss": 0.0085, "num_tokens": 2189722112.0, "reward": 0.095703125, "reward_std": 0.13055095076560974, "rewards/accuracy_reward/mean": 0.095703125, "rewards/accuracy_reward/std": 0.2944713830947876, "step": 2307 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.67578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2030.0, "completions/mean_length": 1555.216796875, "completions/mean_terminated_length": 528.09033203125, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.7879149953059656, "frac_reward_zero_std": 0.75, "grad_norm": 88.92295878105215, "learning_rate": 2.1836105252398483e-07, "loss": 0.0263, "num_tokens": 2190596991.0, "reward": 0.033203125, "reward_std": 0.08082009106874466, "rewards/accuracy_reward/mean": 0.033203125, "rewards/accuracy_reward/std": 0.17934183776378632, "step": 2308 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.748046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1990.0, "completions/mean_length": 1691.08984375, "completions/mean_terminated_length": 631.4263305664062, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7882563796193565, "frac_reward_zero_std": 0.71875, "grad_norm": 74.41291579298706, "learning_rate": 2.1799892212220745e-07, "loss": 0.0093, "num_tokens": 2191538669.0, "reward": 0.068359375, "reward_std": 0.11104501038789749, "rewards/accuracy_reward/mean": 0.068359375, "rewards/accuracy_reward/std": 0.25260838866233826, "step": 2309 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1717.0, "completions/mean_length": 1595.52734375, "completions/mean_terminated_length": 562.9615478515625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7885977639327473, "frac_reward_zero_std": 0.6875, "grad_norm": 28.333364814729112, "learning_rate": 2.1763726293523642e-07, "loss": 0.0204, "num_tokens": 2192430171.0, "reward": 0.080078125, "reward_std": 0.11845915019512177, "rewards/accuracy_reward/mean": 0.080078125, "rewards/accuracy_reward/std": 0.271679550409317, "step": 2310 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.720703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1990.0, "completions/mean_length": 1644.87109375, "completions/mean_terminated_length": 604.6293334960938, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7889391482461381, "frac_reward_zero_std": 0.78125, "grad_norm": 123.15233850278845, "learning_rate": 2.1727607547638073e-07, "loss": 0.0035, "num_tokens": 2193348697.0, "reward": 0.0703125, "reward_std": 0.08126020431518555, "rewards/accuracy_reward/mean": 0.0703125, "rewards/accuracy_reward/std": 0.25592297315597534, "step": 2311 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.740234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1940.0, "completions/mean_length": 1680.373046875, "completions/mean_terminated_length": 632.7744750976562, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.7892805325595289, "frac_reward_zero_std": 0.84375, "grad_norm": 50.89979454443745, "learning_rate": 2.1691536025827982e-07, "loss": -0.0095, "num_tokens": 2194291704.0, "reward": 0.033203125, "reward_std": 0.05628519132733345, "rewards/accuracy_reward/mean": 0.033203125, "rewards/accuracy_reward/std": 0.17934183776378632, "step": 2312 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.748046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1948.0, "completions/mean_length": 1672.73046875, "completions/mean_terminated_length": 558.5581665039062, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7896219168729197, "frac_reward_zero_std": 0.875, "grad_norm": 106.32848149359123, "learning_rate": 2.1655511779290285e-07, "loss": 0.0014, "num_tokens": 2195229534.0, "reward": 0.01953125, "reward_std": 0.04192390665411949, "rewards/accuracy_reward/mean": 0.01953125, "rewards/accuracy_reward/std": 0.1385180652141571, "step": 2313 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.78125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2017.0, "completions/mean_length": 1740.400390625, "completions/mean_terminated_length": 641.8303833007812, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7899633011863105, "frac_reward_zero_std": 0.78125, "grad_norm": 33.28195273694159, "learning_rate": 2.161953485915483e-07, "loss": 0.0168, "num_tokens": 2196196347.0, "reward": 0.07421875, "reward_std": 0.09586012363433838, "rewards/accuracy_reward/mean": 0.07421875, "rewards/accuracy_reward/std": 0.2623828947544098, "step": 2314 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1859.0, "completions/mean_length": 1637.08203125, "completions/mean_terminated_length": 586.9583129882812, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.7903046854997013, "frac_reward_zero_std": 0.8125, "grad_norm": 5.026846147922716, "learning_rate": 2.1583605316484286e-07, "loss": 0.0075, "num_tokens": 2197105093.0, "reward": 0.056640625, "reward_std": 0.08130794763565063, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 2315 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1981.0, "completions/mean_length": 1832.8359375, "completions/mean_terminated_length": 796.1364135742188, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.790646069813092, "frac_reward_zero_std": 0.78125, "grad_norm": 52.12857852157952, "learning_rate": 2.1547723202274039e-07, "loss": 0.0142, "num_tokens": 2198121121.0, "reward": 0.029296875, "reward_std": 0.07014618813991547, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 2316 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.716796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1909.0, "completions/mean_length": 1651.078125, "completions/mean_terminated_length": 646.4552001953125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.7909874541264829, "frac_reward_zero_std": 0.53125, "grad_norm": 16.692409896267765, "learning_rate": 2.1511888567452224e-07, "loss": 0.0215, "num_tokens": 2199041225.0, "reward": 0.05078125, "reward_std": 0.1443656086921692, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 2317 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.681640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2027.0, "completions/mean_length": 1592.9921875, "completions/mean_terminated_length": 619.0858764648438, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.7913288384398737, "frac_reward_zero_std": 0.6875, "grad_norm": 73.11480541739448, "learning_rate": 2.1476101462879504e-07, "loss": 0.0131, "num_tokens": 2199936213.0, "reward": 0.07421875, "reward_std": 0.11482083797454834, "rewards/accuracy_reward/mean": 0.07421875, "rewards/accuracy_reward/std": 0.2623828947544098, "step": 2318 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.806640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1903.0, "completions/mean_length": 1792.884765625, "completions/mean_terminated_length": 728.6161499023438, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7916702227532645, "frac_reward_zero_std": 0.84375, "grad_norm": 32.51115316340324, "learning_rate": 2.1440361939349166e-07, "loss": 0.0078, "num_tokens": 2200935258.0, "reward": 0.01953125, "reward_std": 0.050948239862918854, "rewards/accuracy_reward/mean": 0.02016128972172737, "rewards/accuracy_reward/std": 0.14069372415542603, "step": 2319 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2009.0, "completions/mean_length": 1695.857421875, "completions/mean_terminated_length": 593.991943359375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7920116070666553, "frac_reward_zero_std": 0.75, "grad_norm": 14.80027609902641, "learning_rate": 2.1404670047586905e-07, "loss": 0.0155, "num_tokens": 2201868705.0, "reward": 0.044921875, "reward_std": 0.09126890450716019, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 2320 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.728515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2012.0, "completions/mean_length": 1672.931640625, "completions/mean_terminated_length": 666.4532470703125, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.7923529913800461, "frac_reward_zero_std": 0.71875, "grad_norm": 11.825715894944935, "learning_rate": 2.1369025838250847e-07, "loss": 0.0277, "num_tokens": 2202803022.0, "reward": 0.07421875, "reward_std": 0.12097585201263428, "rewards/accuracy_reward/mean": 0.07421875, "rewards/accuracy_reward/std": 0.2623828947544098, "step": 2321 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.794921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1735.0, "completions/mean_length": 1759.318359375, "completions/mean_terminated_length": 640.3333740234375, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.7926943756934369, "frac_reward_zero_std": 0.84375, "grad_norm": 23.739636668733542, "learning_rate": 2.1333429361931412e-07, "loss": 0.0195, "num_tokens": 2203780081.0, "reward": 0.033203125, "reward_std": 0.06018522381782532, "rewards/accuracy_reward/mean": 0.033203125, "rewards/accuracy_reward/std": 0.17934183776378632, "step": 2322 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1982.0, "completions/mean_length": 1642.646484375, "completions/mean_terminated_length": 645.6959838867188, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "epoch": 0.7930357600068277, "frac_reward_zero_std": 0.78125, "grad_norm": 1.9998989742191888, "learning_rate": 2.1297880669151315e-07, "loss": 0.008, "num_tokens": 2204691932.0, "reward": 0.052734375, "reward_std": 0.06997986882925034, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 2323 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.759765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2037.0, "completions/mean_length": 1713.384765625, "completions/mean_terminated_length": 655.1300659179688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7933771443202184, "frac_reward_zero_std": 0.8125, "grad_norm": 59.33059729153684, "learning_rate": 2.1262379810365404e-07, "loss": -0.0015, "num_tokens": 2205647505.0, "reward": 0.064453125, "reward_std": 0.07737371325492859, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 2324 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.775390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1905.0, "completions/mean_length": 1730.0234375, "completions/mean_terminated_length": 632.31298828125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7937185286336093, "frac_reward_zero_std": 0.78125, "grad_norm": 102.86449558097105, "learning_rate": 2.122692683596064e-07, "loss": 0.0153, "num_tokens": 2206608493.0, "reward": 0.048828125, "reward_std": 0.06371183693408966, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 2325 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.798828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1832.0, "completions/mean_length": 1763.619140625, "completions/mean_terminated_length": 634.378662109375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.7940599129470001, "frac_reward_zero_std": 0.75, "grad_norm": 16.178142510863946, "learning_rate": 2.1191521796256067e-07, "loss": 0.018, "num_tokens": 2207588810.0, "reward": 0.08203125, "reward_std": 0.09629231691360474, "rewards/accuracy_reward/mean": 0.08203125, "rewards/accuracy_reward/std": 0.2746807038784027, "step": 2326 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.724609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1869.0, "completions/mean_length": 1652.4609375, "completions/mean_terminated_length": 611.71630859375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7944012972603909, "frac_reward_zero_std": 0.65625, "grad_norm": 69.64055666030006, "learning_rate": 2.1156164741502639e-07, "loss": 0.0146, "num_tokens": 2208508854.0, "reward": 0.095703125, "reward_std": 0.136918306350708, "rewards/accuracy_reward/mean": 0.095703125, "rewards/accuracy_reward/std": 0.2944713830947876, "step": 2327 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1853.0, "completions/mean_length": 1742.3046875, "completions/mean_terminated_length": 598.7777709960938, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7947426815737817, "frac_reward_zero_std": 0.78125, "grad_norm": 21.912850176433427, "learning_rate": 2.1120855721883253e-07, "loss": 0.0021, "num_tokens": 2209486978.0, "reward": 0.0390625, "reward_std": 0.07289456576108932, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 2328 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.685546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1980.0, "completions/mean_length": 1561.826171875, "completions/mean_terminated_length": 501.9068298339844, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7950840658871725, "frac_reward_zero_std": 0.71875, "grad_norm": 12.448221811656028, "learning_rate": 2.1085594787512579e-07, "loss": 0.0296, "num_tokens": 2210368569.0, "reward": 0.0625, "reward_std": 0.1045791357755661, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.2422981858253479, "step": 2329 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.775390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2038.0, "completions/mean_length": 1725.544921875, "completions/mean_terminated_length": 612.3739013671875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7954254502005633, "frac_reward_zero_std": 0.6875, "grad_norm": 54.44276330759143, "learning_rate": 2.105038198843707e-07, "loss": 0.008, "num_tokens": 2211318880.0, "reward": 0.0546875, "reward_std": 0.11587189882993698, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 2330 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.716796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1914.0, "completions/mean_length": 1633.37890625, "completions/mean_terminated_length": 583.9586181640625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7957668345139541, "frac_reward_zero_std": 0.8125, "grad_norm": 13.101626387600385, "learning_rate": 2.101521737463487e-07, "loss": 0.0254, "num_tokens": 2212234402.0, "reward": 0.021484375, "reward_std": 0.059305962175130844, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 2331 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.76953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1839.0, "completions/mean_length": 1721.04296875, "completions/mean_terminated_length": 629.3389892578125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7961082188273448, "frac_reward_zero_std": 0.9375, "grad_norm": 10.763200878778521, "learning_rate": 2.0980100996015694e-07, "loss": -0.0012, "num_tokens": 2213204824.0, "reward": 0.00390625, "reward_std": 0.015625, "rewards/accuracy_reward/mean": 0.00390625, "rewards/accuracy_reward/std": 0.06243881583213806, "step": 2332 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2002.0, "completions/mean_length": 1694.966796875, "completions/mean_terminated_length": 590.3145141601562, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.7964496031407357, "frac_reward_zero_std": 0.9375, "grad_norm": 0.21259084836439932, "learning_rate": 2.094503290242084e-07, "loss": 0.003, "num_tokens": 2214156023.0, "reward": 0.02734375, "reward_std": 0.024649331346154213, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 2333 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.759765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1877.0, "completions/mean_length": 1706.279296875, "completions/mean_terminated_length": 625.5527954101562, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7967909874541265, "frac_reward_zero_std": 0.84375, "grad_norm": 52.74611986954134, "learning_rate": 2.091001314362303e-07, "loss": -0.0056, "num_tokens": 2215105702.0, "reward": 0.025390625, "reward_std": 0.052045635879039764, "rewards/accuracy_reward/mean": 0.025390625, "rewards/accuracy_reward/std": 0.15746226906776428, "step": 2334 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.716796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1901.0, "completions/mean_length": 1632.55078125, "completions/mean_terminated_length": 581.0344848632812, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7971323717675173, "frac_reward_zero_std": 0.84375, "grad_norm": 85.39331085513942, "learning_rate": 2.087504176932643e-07, "loss": -0.0087, "num_tokens": 2216013488.0, "reward": 0.048828125, "reward_std": 0.06391072273254395, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 2335 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.818359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1947.0, "completions/mean_length": 1760.265625, "completions/mean_terminated_length": 463.9139709472656, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.7974737560809081, "frac_reward_zero_std": 0.84375, "grad_norm": 7.826832074826158, "learning_rate": 2.0840118829166498e-07, "loss": 0.0072, "num_tokens": 2216994296.0, "reward": 0.033203125, "reward_std": 0.05739613622426987, "rewards/accuracy_reward/mean": 0.033203125, "rewards/accuracy_reward/std": 0.17934183776378632, "step": 2336 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.724609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2022.0, "completions/mean_length": 1688.720703125, "completions/mean_terminated_length": 743.3829345703125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7978151403942989, "frac_reward_zero_std": 0.75, "grad_norm": 82.417618219873, "learning_rate": 2.0805244372709952e-07, "loss": 0.0046, "num_tokens": 2217934457.0, "reward": 0.08203125, "reward_std": 0.10360593348741531, "rewards/accuracy_reward/mean": 0.08203125, "rewards/accuracy_reward/std": 0.2746807038784027, "step": 2337 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.75, "completions/max_length": 2048.0, "completions/max_terminated_length": 1976.0, "completions/mean_length": 1704.603515625, "completions/mean_terminated_length": 674.4140625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7981565247076897, "frac_reward_zero_std": 0.71875, "grad_norm": 15.470507158949687, "learning_rate": 2.077041844945472e-07, "loss": 0.0228, "num_tokens": 2218884110.0, "reward": 0.03515625, "reward_std": 0.08989076316356659, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 2338 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.77734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1883.0, "completions/mean_length": 1717.638671875, "completions/mean_terminated_length": 564.2719116210938, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.7984979090210805, "frac_reward_zero_std": 0.875, "grad_norm": 5.885528113056802, "learning_rate": 2.0735641108829813e-07, "loss": 0.0013, "num_tokens": 2219843541.0, "reward": 0.02734375, "reward_std": 0.05032937601208687, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 2339 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.806640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1935.0, "completions/mean_length": 1782.3984375, "completions/mean_terminated_length": 674.80810546875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.7988392933344712, "frac_reward_zero_std": 0.75, "grad_norm": 37.54077045209394, "learning_rate": 2.070091240019533e-07, "loss": 0.0038, "num_tokens": 2220832177.0, "reward": 0.06640625, "reward_std": 0.0907755047082901, "rewards/accuracy_reward/mean": 0.06640625, "rewards/accuracy_reward/std": 0.2492343932390213, "step": 2340 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.783203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2015.0, "completions/mean_length": 1727.375, "completions/mean_terminated_length": 569.0811157226562, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.799180677647862, "frac_reward_zero_std": 0.84375, "grad_norm": 44.208385368551035, "learning_rate": 2.066623237284229e-07, "loss": 0.0051, "num_tokens": 2221787137.0, "reward": 0.01953125, "reward_std": 0.050948236137628555, "rewards/accuracy_reward/mean": 0.02016128972172737, "rewards/accuracy_reward/std": 0.14069372415542603, "step": 2341 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1674.53515625, "completions/mean_terminated_length": 642.0147094726562, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7995220619612529, "frac_reward_zero_std": 0.75, "grad_norm": 47.00040401198957, "learning_rate": 2.0631601075992677e-07, "loss": -0.0165, "num_tokens": 2222725283.0, "reward": 0.064453125, "reward_std": 0.09673243016004562, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 2342 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2026.0, "completions/mean_length": 1666.021484375, "completions/mean_terminated_length": 689.8541870117188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.7998634462746437, "frac_reward_zero_std": 0.75, "grad_norm": 32.50472099380991, "learning_rate": 2.0597018558799272e-07, "loss": 0.0072, "num_tokens": 2223658622.0, "reward": 0.04296875, "reward_std": 0.09028453379869461, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 2343 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.77734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1717.0, "completions/mean_length": 1733.44921875, "completions/mean_terminated_length": 635.2807006835938, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.8002048305880345, "frac_reward_zero_std": 0.875, "grad_norm": 20.74382005587314, "learning_rate": 2.056248487034562e-07, "loss": 0.0184, "num_tokens": 2224625332.0, "reward": 0.033203125, "reward_std": 0.03411140665411949, "rewards/accuracy_reward/mean": 0.033203125, "rewards/accuracy_reward/std": 0.17934183776378632, "step": 2344 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.73046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1928.0, "completions/mean_length": 1655.8359375, "completions/mean_terminated_length": 593.0145263671875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8005462149014253, "frac_reward_zero_std": 0.71875, "grad_norm": 74.26744304555125, "learning_rate": 2.0528000059645995e-07, "loss": 0.001, "num_tokens": 2225555552.0, "reward": 0.09765625, "reward_std": 0.11478663980960846, "rewards/accuracy_reward/mean": 0.09765625, "rewards/accuracy_reward/std": 0.29713961482048035, "step": 2345 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2003.0, "completions/mean_length": 1638.267578125, "completions/mean_terminated_length": 458.73486328125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8008875992148161, "frac_reward_zero_std": 0.84375, "grad_norm": 21.947734292105586, "learning_rate": 2.0493564175645256e-07, "loss": 0.0117, "num_tokens": 2226482169.0, "reward": 0.044921875, "reward_std": 0.06417001038789749, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 2346 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.8125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1790.0, "completions/mean_length": 1782.31640625, "completions/mean_terminated_length": 631.6041870117188, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.8012289835282069, "frac_reward_zero_std": 0.78125, "grad_norm": 157.2591046553519, "learning_rate": 2.0459177267218878e-07, "loss": 0.0157, "num_tokens": 2227472123.0, "reward": 0.037109375, "reward_std": 0.07355976849794388, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 2347 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.75390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2019.0, "completions/mean_length": 1715.0, "completions/mean_terminated_length": 694.857177734375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8015703678415976, "frac_reward_zero_std": 0.78125, "grad_norm": 36.2203498230152, "learning_rate": 2.042483938317276e-07, "loss": 0.0135, "num_tokens": 2228425995.0, "reward": 0.0390625, "reward_std": 0.07368633151054382, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 2348 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.78125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2003.0, "completions/mean_length": 1732.705078125, "completions/mean_terminated_length": 606.6517944335938, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.8019117521549884, "frac_reward_zero_std": 0.78125, "grad_norm": 4.299639884149655, "learning_rate": 2.0390550572243242e-07, "loss": 0.0214, "num_tokens": 2229392084.0, "reward": 0.03125, "reward_std": 0.07394562661647797, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 2349 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.78515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1728.14453125, "completions/mean_terminated_length": 559.2181396484375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.8022531364683793, "frac_reward_zero_std": 0.75, "grad_norm": 112.52357578693962, "learning_rate": 2.0356310883097045e-07, "loss": 0.0141, "num_tokens": 2230350286.0, "reward": 0.048828125, "reward_std": 0.09038965404033661, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 2350 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.740234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1910.0, "completions/mean_length": 1669.35546875, "completions/mean_terminated_length": 590.3609008789062, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.8025945207817701, "frac_reward_zero_std": 0.8125, "grad_norm": 35.321963871602684, "learning_rate": 2.0322120364331119e-07, "loss": 0.0159, "num_tokens": 2231290308.0, "reward": 0.052734375, "reward_std": 0.07669496536254883, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 2351 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.72265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1941.0, "completions/mean_length": 1617.904296875, "completions/mean_terminated_length": 497.2323913574219, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8029359050951609, "frac_reward_zero_std": 0.75, "grad_norm": 16.453376373906206, "learning_rate": 2.028797906447268e-07, "loss": 0.0175, "num_tokens": 2232200227.0, "reward": 0.064453125, "reward_std": 0.09271937608718872, "rewards/accuracy_reward/mean": 0.06653226166963577, "rewards/accuracy_reward/std": 0.2494617998600006, "step": 2352 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.775390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1788.0, "completions/mean_length": 1701.537109375, "completions/mean_terminated_length": 505.4869384765625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8032772894085517, "frac_reward_zero_std": 0.78125, "grad_norm": 102.45565549040242, "learning_rate": 2.025388703197903e-07, "loss": -0.0068, "num_tokens": 2233151638.0, "reward": 0.0546875, "reward_std": 0.07795868813991547, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 2353 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.720703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1985.0, "completions/mean_length": 1631.576171875, "completions/mean_terminated_length": 557.0279541015625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.8036186737219425, "frac_reward_zero_std": 0.65625, "grad_norm": 116.76885641439672, "learning_rate": 2.0219844315237595e-07, "loss": 0.0105, "num_tokens": 2234070349.0, "reward": 0.0390625, "reward_std": 0.1049705445766449, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 2354 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1991.0, "completions/mean_length": 1633.908203125, "completions/mean_terminated_length": 533.6071166992188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8039600580353333, "frac_reward_zero_std": 0.78125, "grad_norm": 21.94041085031215, "learning_rate": 2.0185850962565782e-07, "loss": 0.0124, "num_tokens": 2234989886.0, "reward": 0.037109375, "reward_std": 0.07003315538167953, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 2355 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.775390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1933.0, "completions/mean_length": 1716.90234375, "completions/mean_terminated_length": 573.8956298828125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.804301442348724, "frac_reward_zero_std": 0.6875, "grad_norm": 33.16970325150947, "learning_rate": 2.0151907022210917e-07, "loss": 0.0109, "num_tokens": 2235942428.0, "reward": 0.09765625, "reward_std": 0.11564444750547409, "rewards/accuracy_reward/mean": 0.09765625, "rewards/accuracy_reward/std": 0.29713961482048035, "step": 2356 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.818359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2015.0, "completions/mean_length": 1829.171875, "completions/mean_terminated_length": 843.8709716796875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.8046428266621148, "frac_reward_zero_std": 0.78125, "grad_norm": 57.933791522022574, "learning_rate": 2.0118012542350245e-07, "loss": -0.005, "num_tokens": 2236963508.0, "reward": 0.03515625, "reward_std": 0.07477159798145294, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 2357 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.724609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2014.0, "completions/mean_length": 1683.123046875, "completions/mean_terminated_length": 723.0567016601562, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.8049842109755057, "frac_reward_zero_std": 0.65625, "grad_norm": 47.865891343447345, "learning_rate": 2.0084167571090753e-07, "loss": 0.0175, "num_tokens": 2237913763.0, "reward": 0.064453125, "reward_std": 0.12451459467411041, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 2358 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.77734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1971.0, "completions/mean_length": 1732.88671875, "completions/mean_terminated_length": 632.75439453125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8053255952888965, "frac_reward_zero_std": 0.84375, "grad_norm": 14.918082991180457, "learning_rate": 2.0050372156469214e-07, "loss": 0.0016, "num_tokens": 2238882521.0, "reward": 0.009765625, "reward_std": 0.0390625, "rewards/accuracy_reward/mean": 0.009765625, "rewards/accuracy_reward/std": 0.09843364357948303, "step": 2359 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2028.0, "completions/mean_length": 1695.150390625, "completions/mean_terminated_length": 719.625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.8056669796022873, "frac_reward_zero_std": 0.78125, "grad_norm": 41.19709490896527, "learning_rate": 2.0016626346452027e-07, "loss": 0.005, "num_tokens": 2239820374.0, "reward": 0.05078125, "reward_std": 0.08477334678173065, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 2360 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.69921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2003.0, "completions/mean_length": 1632.69140625, "completions/mean_terminated_length": 667.2337646484375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.8060083639156781, "frac_reward_zero_std": 0.6875, "grad_norm": 20.18628436938063, "learning_rate": 1.998293018893518e-07, "loss": 0.0205, "num_tokens": 2240727432.0, "reward": 0.0703125, "reward_std": 0.10975530743598938, "rewards/accuracy_reward/mean": 0.0703125, "rewards/accuracy_reward/std": 0.25592297315597534, "step": 2361 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.75, "completions/max_length": 2048.0, "completions/max_terminated_length": 1687.0, "completions/mean_length": 1674.759765625, "completions/mean_terminated_length": 555.0390625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.8063497482290689, "frac_reward_zero_std": 0.5625, "grad_norm": 116.11114513473144, "learning_rate": 1.994928373174423e-07, "loss": 0.0221, "num_tokens": 2241666797.0, "reward": 0.103515625, "reward_std": 0.15234442055225372, "rewards/accuracy_reward/mean": 0.103515625, "rewards/accuracy_reward/std": 0.30492907762527466, "step": 2362 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.771484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1760.0, "completions/mean_length": 1740.38671875, "completions/mean_terminated_length": 701.86328125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8066911325424597, "frac_reward_zero_std": 0.6875, "grad_norm": 24.14525121534884, "learning_rate": 1.991568702263415e-07, "loss": 0.0072, "num_tokens": 2242635955.0, "reward": 0.060546875, "reward_std": 0.10712136328220367, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 2363 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.66796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1816.0, "completions/mean_length": 1559.375, "completions/mean_terminated_length": 576.37646484375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.8070325168558504, "frac_reward_zero_std": 0.6875, "grad_norm": 71.130220900969, "learning_rate": 1.9882140109289342e-07, "loss": 0.013, "num_tokens": 2243504547.0, "reward": 0.0625, "reward_std": 0.11211783438920975, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.2422981858253479, "step": 2364 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.806640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1960.0, "completions/mean_length": 1788.83203125, "completions/mean_terminated_length": 707.6565551757812, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.8073739011692412, "frac_reward_zero_std": 0.8125, "grad_norm": 22.30202303856115, "learning_rate": 1.9848643039323499e-07, "loss": 0.0057, "num_tokens": 2244499325.0, "reward": 0.029296875, "reward_std": 0.06271954625844955, "rewards/accuracy_reward/mean": 0.030241934582591057, "rewards/accuracy_reward/std": 0.1714252382516861, "step": 2365 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.74609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1962.0, "completions/mean_length": 1694.49609375, "completions/mean_terminated_length": 655.7384643554688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.807715285482632, "frac_reward_zero_std": 0.75, "grad_norm": 50.665211144706575, "learning_rate": 1.9815195860279594e-07, "loss": 0.0044, "num_tokens": 2245448715.0, "reward": 0.072265625, "reward_std": 0.10057821869850159, "rewards/accuracy_reward/mean": 0.072265625, "rewards/accuracy_reward/std": 0.2591804563999176, "step": 2366 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.779296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1940.0, "completions/mean_length": 1733.873046875, "completions/mean_terminated_length": 625.1150512695312, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8080566697960229, "frac_reward_zero_std": 0.8125, "grad_norm": 27.702283829907472, "learning_rate": 1.978179861962978e-07, "loss": -0.0013, "num_tokens": 2246422058.0, "reward": 0.025390625, "reward_std": 0.06041031330823898, "rewards/accuracy_reward/mean": 0.025390625, "rewards/accuracy_reward/std": 0.15746226906776428, "step": 2367 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.75, "completions/max_length": 2048.0, "completions/max_terminated_length": 1906.0, "completions/mean_length": 1666.736328125, "completions/mean_terminated_length": 522.9453125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8083980541094137, "frac_reward_zero_std": 0.78125, "grad_norm": 19.522280171389955, "learning_rate": 1.9748451364775318e-07, "loss": 0.0219, "num_tokens": 2247358627.0, "reward": 0.0546875, "reward_std": 0.08408090472221375, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 2368 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.67578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1943.0, "completions/mean_length": 1575.283203125, "completions/mean_terminated_length": 589.98193359375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8087394384228045, "frac_reward_zero_std": 0.65625, "grad_norm": 73.70984172439701, "learning_rate": 1.9715154143046558e-07, "loss": 0.0295, "num_tokens": 2248240788.0, "reward": 0.09765625, "reward_std": 0.13566282391548157, "rewards/accuracy_reward/mean": 0.09765625, "rewards/accuracy_reward/std": 0.29713961482048035, "step": 2369 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.666015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1996.0, "completions/mean_length": 1572.32421875, "completions/mean_terminated_length": 623.75439453125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8090808227361953, "frac_reward_zero_std": 0.8125, "grad_norm": 53.90882983180562, "learning_rate": 1.968190700170279e-07, "loss": 0.009, "num_tokens": 2249119210.0, "reward": 0.0390625, "reward_std": 0.072782501578331, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 2370 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.751953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2003.0, "completions/mean_length": 1687.39453125, "completions/mean_terminated_length": 594.55908203125, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.8094222070495861, "frac_reward_zero_std": 0.65625, "grad_norm": 69.25571043436973, "learning_rate": 1.9648709987932282e-07, "loss": -0.0103, "num_tokens": 2250075780.0, "reward": 0.076171875, "reward_std": 0.13715851306915283, "rewards/accuracy_reward/mean": 0.076171875, "rewards/accuracy_reward/std": 0.26553234457969666, "step": 2371 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2024.0, "completions/mean_length": 1623.03125, "completions/mean_terminated_length": 537.0, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.8097635913629768, "frac_reward_zero_std": 0.875, "grad_norm": 28.3392348422541, "learning_rate": 1.9615563148852092e-07, "loss": 0.0008, "num_tokens": 2250988084.0, "reward": 0.021484375, "reward_std": 0.043847277760505676, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 2372 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1910.0, "completions/mean_length": 1686.654296875, "completions/mean_terminated_length": 646.4166870117188, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8101049756763676, "frac_reward_zero_std": 0.8125, "grad_norm": 32.84198159466664, "learning_rate": 1.9582466531508135e-07, "loss": 0.0172, "num_tokens": 2251933747.0, "reward": 0.072265625, "reward_std": 0.06975477933883667, "rewards/accuracy_reward/mean": 0.072265625, "rewards/accuracy_reward/std": 0.2591804563999176, "step": 2373 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1879.0, "completions/mean_length": 1730.986328125, "completions/mean_terminated_length": 648.7672119140625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.8104463599897584, "frac_reward_zero_std": 0.78125, "grad_norm": 69.37044844794951, "learning_rate": 1.9549420182874956e-07, "loss": -0.0041, "num_tokens": 2252900156.0, "reward": 0.03515625, "reward_std": 0.07630910724401474, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 2374 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.77734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1983.0, "completions/mean_length": 1760.072265625, "completions/mean_terminated_length": 754.8508911132812, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8107877443031493, "frac_reward_zero_std": 0.78125, "grad_norm": 43.44709629053883, "learning_rate": 1.9516424149855829e-07, "loss": 0.0119, "num_tokens": 2253877697.0, "reward": 0.056640625, "reward_std": 0.08930579572916031, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 2375 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.716796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2019.0, "completions/mean_length": 1638.06640625, "completions/mean_terminated_length": 600.8482666015625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8111291286165401, "frac_reward_zero_std": 0.75, "grad_norm": 40.367098943748786, "learning_rate": 1.94834784792826e-07, "loss": 0.0058, "num_tokens": 2254795859.0, "reward": 0.0546875, "reward_std": 0.08587866276502609, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 2376 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.822265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1643.0, "completions/mean_length": 1773.3125, "completions/mean_terminated_length": 502.5054931640625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8114705129299309, "frac_reward_zero_std": 0.78125, "grad_norm": 23.0068042387084, "learning_rate": 1.9450583217915595e-07, "loss": 0.0083, "num_tokens": 2255784499.0, "reward": 0.0546875, "reward_std": 0.07961063086986542, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 2377 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1991.0, "completions/mean_length": 1614.546875, "completions/mean_terminated_length": 587.9473876953125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8118118972433217, "frac_reward_zero_std": 0.75, "grad_norm": 113.46613152516082, "learning_rate": 1.9417738412443647e-07, "loss": 0.0177, "num_tokens": 2256685483.0, "reward": 0.0390625, "reward_std": 0.08604402840137482, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 2378 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.76171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1966.0, "completions/mean_length": 1707.02734375, "completions/mean_terminated_length": 617.0327758789062, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8121532815567125, "frac_reward_zero_std": 0.8125, "grad_norm": 11.710617153134043, "learning_rate": 1.938494410948394e-07, "loss": 0.0159, "num_tokens": 2257635721.0, "reward": 0.03515625, "reward_std": 0.06948098540306091, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 2379 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.78515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1998.0, "completions/mean_length": 1759.46484375, "completions/mean_terminated_length": 705.0, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8124946658701032, "frac_reward_zero_std": 0.75, "grad_norm": 31.73394419984976, "learning_rate": 1.9352200355581988e-07, "loss": 0.0014, "num_tokens": 2258610215.0, "reward": 0.08984375, "reward_std": 0.08851956576108932, "rewards/accuracy_reward/mean": 0.0927419364452362, "rewards/accuracy_reward/std": 0.2903633117675781, "step": 2380 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.708984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2034.0, "completions/mean_length": 1630.16796875, "completions/mean_terminated_length": 612.2282104492188, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.812836050183494, "frac_reward_zero_std": 0.78125, "grad_norm": 7.5451585846974485, "learning_rate": 1.9319507197211538e-07, "loss": 0.024, "num_tokens": 2259521261.0, "reward": 0.037109375, "reward_std": 0.07729348540306091, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 2381 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2005.0, "completions/mean_length": 1633.615234375, "completions/mean_terminated_length": 594.8150634765625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8131774344968848, "frac_reward_zero_std": 0.78125, "grad_norm": 10.269616327278692, "learning_rate": 1.9286864680774578e-07, "loss": 0.0121, "num_tokens": 2260441176.0, "reward": 0.025390625, "reward_std": 0.06755761802196503, "rewards/accuracy_reward/mean": 0.026209676638245583, "rewards/accuracy_reward/std": 0.1599196344614029, "step": 2382 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.736328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2020.0, "completions/mean_length": 1655.54296875, "completions/mean_terminated_length": 559.5703735351562, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8135188188102757, "frac_reward_zero_std": 0.65625, "grad_norm": 86.46369488355815, "learning_rate": 1.9254272852601193e-07, "loss": 0.0066, "num_tokens": 2261361374.0, "reward": 0.095703125, "reward_std": 0.127755269408226, "rewards/accuracy_reward/mean": 0.095703125, "rewards/accuracy_reward/std": 0.2944713830947876, "step": 2383 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.75390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1991.0, "completions/mean_length": 1704.91796875, "completions/mean_terminated_length": 653.888916015625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8138602031236665, "frac_reward_zero_std": 0.8125, "grad_norm": 86.507797004118, "learning_rate": 1.9221731758949498e-07, "loss": 0.0059, "num_tokens": 2262311892.0, "reward": 0.05859375, "reward_std": 0.08798330277204514, "rewards/accuracy_reward/mean": 0.05859375, "rewards/accuracy_reward/std": 0.23509246110916138, "step": 2384 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.75, "completions/max_length": 2048.0, "completions/max_terminated_length": 2007.0, "completions/mean_length": 1717.48828125, "completions/mean_terminated_length": 725.953125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8142015874370573, "frac_reward_zero_std": 0.71875, "grad_norm": 36.90936366866858, "learning_rate": 1.918924144600566e-07, "loss": 0.011, "num_tokens": 2263277118.0, "reward": 0.0546875, "reward_std": 0.10571042448282242, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 2385 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2039.0, "completions/mean_length": 1619.349609375, "completions/mean_terminated_length": 676.3187866210938, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.8145429717504481, "frac_reward_zero_std": 0.8125, "grad_norm": 59.32024847477373, "learning_rate": 1.915680195988369e-07, "loss": -0.0146, "num_tokens": 2264187777.0, "reward": 0.03125, "reward_std": 0.06469620764255524, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 2386 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.669921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1930.0, "completions/mean_length": 1562.865234375, "completions/mean_terminated_length": 578.2426147460938, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8148843560638389, "frac_reward_zero_std": 0.84375, "grad_norm": 36.401452128281754, "learning_rate": 1.912441334662554e-07, "loss": -0.0007, "num_tokens": 2265067804.0, "reward": 0.048828125, "reward_std": 0.0673912912607193, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 2387 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.748046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2005.0, "completions/mean_length": 1674.37890625, "completions/mean_terminated_length": 565.534912109375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8152257403772296, "frac_reward_zero_std": 0.65625, "grad_norm": 37.00439333912248, "learning_rate": 1.9092075652200894e-07, "loss": -0.0073, "num_tokens": 2266010782.0, "reward": 0.0390625, "reward_std": 0.104032501578331, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 2388 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2032.0, "completions/mean_length": 1676.6875, "completions/mean_terminated_length": 650.11767578125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.8155671246906204, "frac_reward_zero_std": 0.78125, "grad_norm": 7.940244333587208, "learning_rate": 1.9059788922507213e-07, "loss": 0.0054, "num_tokens": 2266946702.0, "reward": 0.041015625, "reward_std": 0.06943464279174805, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 2389 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.716796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1804.0, "completions/mean_length": 1648.38671875, "completions/mean_terminated_length": 636.9517211914062, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8159085090040112, "frac_reward_zero_std": 0.84375, "grad_norm": 73.98209700648937, "learning_rate": 1.902755320336961e-07, "loss": 0.0003, "num_tokens": 2267868324.0, "reward": 0.041015625, "reward_std": 0.06234722584486008, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 2390 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1919.0, "completions/mean_length": 1568.408203125, "completions/mean_terminated_length": 586.3869018554688, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.816249893317402, "frac_reward_zero_std": 0.65625, "grad_norm": 31.503703525765093, "learning_rate": 1.899536854054079e-07, "loss": 0.0103, "num_tokens": 2268751061.0, "reward": 0.072265625, "reward_std": 0.12576615810394287, "rewards/accuracy_reward/mean": 0.072265625, "rewards/accuracy_reward/std": 0.2591804563999176, "step": 2391 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.681640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2021.0, "completions/mean_length": 1587.634765625, "completions/mean_terminated_length": 601.9447631835938, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8165912776307929, "frac_reward_zero_std": 0.6875, "grad_norm": 37.65155847365325, "learning_rate": 1.8963234979700986e-07, "loss": 0.0106, "num_tokens": 2269639578.0, "reward": 0.0859375, "reward_std": 0.1249004602432251, "rewards/accuracy_reward/mean": 0.0859375, "rewards/accuracy_reward/std": 0.28054583072662354, "step": 2392 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.73046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1798.0, "completions/mean_length": 1665.451171875, "completions/mean_terminated_length": 628.6884155273438, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.8169326619441837, "frac_reward_zero_std": 0.71875, "grad_norm": 49.74834220852008, "learning_rate": 1.8931152566457903e-07, "loss": 0.0297, "num_tokens": 2270566833.0, "reward": 0.04296875, "reward_std": 0.09424237906932831, "rewards/accuracy_reward/mean": 0.04435483738780022, "rewards/accuracy_reward/std": 0.2060900777578354, "step": 2393 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.65234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1917.0, "completions/mean_length": 1561.248046875, "completions/mean_terminated_length": 647.9044799804688, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.8172740462575745, "frac_reward_zero_std": 0.71875, "grad_norm": 41.887315389048005, "learning_rate": 1.8899121346346682e-07, "loss": 0.0226, "num_tokens": 2271438928.0, "reward": 0.087890625, "reward_std": 0.10735208541154861, "rewards/accuracy_reward/mean": 0.087890625, "rewards/accuracy_reward/std": 0.2834126651287079, "step": 2394 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1843.0, "completions/mean_length": 1603.369140625, "completions/mean_terminated_length": 550.2960815429688, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8176154305709653, "frac_reward_zero_std": 0.78125, "grad_norm": 34.538404230653015, "learning_rate": 1.8867141364829758e-07, "loss": -0.0122, "num_tokens": 2272344381.0, "reward": 0.060546875, "reward_std": 0.08175812661647797, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 2395 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.716796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1980.0, "completions/mean_length": 1653.189453125, "completions/mean_terminated_length": 653.9103393554688, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.817956814884356, "frac_reward_zero_std": 0.84375, "grad_norm": 24.446545873701883, "learning_rate": 1.8835212667296873e-07, "loss": 0.0089, "num_tokens": 2273268622.0, "reward": 0.02734375, "reward_std": 0.05311024561524391, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 2396 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1963.0, "completions/mean_length": 1565.212890625, "completions/mean_terminated_length": 674.7388916015625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8182981991977468, "frac_reward_zero_std": 0.71875, "grad_norm": 35.2135498172077, "learning_rate": 1.8803335299064998e-07, "loss": 0.0131, "num_tokens": 2274145355.0, "reward": 0.048828125, "reward_std": 0.1022823378443718, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 2397 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.666015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1848.0, "completions/mean_length": 1558.78125, "completions/mean_terminated_length": 583.2047119140625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8186395835111376, "frac_reward_zero_std": 0.78125, "grad_norm": 37.83840348933624, "learning_rate": 1.8771509305378186e-07, "loss": -0.0026, "num_tokens": 2275023899.0, "reward": 0.064453125, "reward_std": 0.09268517792224884, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 2398 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.693359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1840.0, "completions/mean_length": 1601.35546875, "completions/mean_terminated_length": 591.4267578125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8189809678245284, "frac_reward_zero_std": 0.8125, "grad_norm": 60.504144028243445, "learning_rate": 1.8739734731407646e-07, "loss": 0.0129, "num_tokens": 2275919105.0, "reward": 0.05078125, "reward_std": 0.06558094918727875, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 2399 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.69140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1949.0, "completions/mean_length": 1625.76953125, "completions/mean_terminated_length": 679.759521484375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8193223521379193, "frac_reward_zero_std": 0.71875, "grad_norm": 22.556367182540573, "learning_rate": 1.870801162225155e-07, "loss": 0.0295, "num_tokens": 2276832283.0, "reward": 0.0546875, "reward_std": 0.08940431475639343, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 2400 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.673828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1945.0, "completions/mean_length": 1600.62109375, "completions/mean_terminated_length": 676.395263671875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8196637364513101, "frac_reward_zero_std": 0.84375, "grad_norm": 41.92500824991513, "learning_rate": 1.8676340022935073e-07, "loss": 0.0217, "num_tokens": 2277723465.0, "reward": 0.0390625, "reward_std": 0.0625, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 2401 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.724609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1907.0, "completions/mean_length": 1632.537109375, "completions/mean_terminated_length": 539.3687744140625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8200051207647009, "frac_reward_zero_std": 0.6875, "grad_norm": 39.20352764063072, "learning_rate": 1.8644719978410227e-07, "loss": 0.047, "num_tokens": 2278633276.0, "reward": 0.080078125, "reward_std": 0.11984425783157349, "rewards/accuracy_reward/mean": 0.080078125, "rewards/accuracy_reward/std": 0.271679550409317, "step": 2402 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1931.0, "completions/mean_length": 1501.068359375, "completions/mean_terminated_length": 619.2805786132812, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8203465050780917, "frac_reward_zero_std": 0.78125, "grad_norm": 27.758081654029525, "learning_rate": 1.861315153355592e-07, "loss": 0.0212, "num_tokens": 2279474719.0, "reward": 0.060546875, "reward_std": 0.08532553166151047, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 2403 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.67578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2038.0, "completions/mean_length": 1594.44140625, "completions/mean_terminated_length": 649.072265625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.8206878893914825, "frac_reward_zero_std": 0.875, "grad_norm": 3.270552634233083, "learning_rate": 1.8581634733177758e-07, "loss": -0.0065, "num_tokens": 2280364929.0, "reward": 0.017578125, "reward_std": 0.043680962175130844, "rewards/accuracy_reward/mean": 0.017578125, "rewards/accuracy_reward/std": 0.13154059648513794, "step": 2404 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1915.0, "completions/mean_length": 1543.373046875, "completions/mean_terminated_length": 545.8546752929688, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8210292737048732, "frac_reward_zero_std": 0.78125, "grad_norm": 40.85661673042326, "learning_rate": 1.8550169622008078e-07, "loss": 0.007, "num_tokens": 2281231232.0, "reward": 0.044921875, "reward_std": 0.07718046009540558, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 2405 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.65625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1792.0, "completions/mean_length": 1543.181640625, "completions/mean_terminated_length": 579.7556762695312, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.821370658018264, "frac_reward_zero_std": 0.71875, "grad_norm": 43.14568712145668, "learning_rate": 1.851875624470586e-07, "loss": 0.0147, "num_tokens": 2282103037.0, "reward": 0.072265625, "reward_std": 0.09926675260066986, "rewards/accuracy_reward/mean": 0.072265625, "rewards/accuracy_reward/std": 0.2591804563999176, "step": 2406 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1973.0, "completions/mean_length": 1572.271484375, "completions/mean_terminated_length": 631.8779296875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8217120423316548, "frac_reward_zero_std": 0.75, "grad_norm": 97.20281769766575, "learning_rate": 1.8487394645856636e-07, "loss": 0.0101, "num_tokens": 2282981304.0, "reward": 0.072265625, "reward_std": 0.0971047431230545, "rewards/accuracy_reward/mean": 0.072265625, "rewards/accuracy_reward/std": 0.2591804563999176, "step": 2407 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.697265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1982.0, "completions/mean_length": 1637.72265625, "completions/mean_terminated_length": 692.7612915039062, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.8220534266450457, "frac_reward_zero_std": 0.75, "grad_norm": 49.18603278315709, "learning_rate": 1.8456084869972472e-07, "loss": 0.0193, "num_tokens": 2283887066.0, "reward": 0.05859375, "reward_std": 0.08325589448213577, "rewards/accuracy_reward/mean": 0.05859375, "rewards/accuracy_reward/std": 0.23509246110916138, "step": 2408 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1922.0, "completions/mean_length": 1602.955078125, "completions/mean_terminated_length": 623.8562622070312, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8223948109584365, "frac_reward_zero_std": 0.78125, "grad_norm": 30.633507567713714, "learning_rate": 1.8424826961491852e-07, "loss": -0.0071, "num_tokens": 2284783571.0, "reward": 0.0546875, "reward_std": 0.08518621325492859, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 2409 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.75, "completions/max_length": 2048.0, "completions/max_terminated_length": 2017.0, "completions/mean_length": 1696.619140625, "completions/mean_terminated_length": 642.4765625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8227361952718273, "frac_reward_zero_std": 0.75, "grad_norm": 13.546852089517008, "learning_rate": 1.8393620964779675e-07, "loss": 0.0071, "num_tokens": 2285725712.0, "reward": 0.029296875, "reward_std": 0.07917051017284393, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 2410 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.716796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2043.0, "completions/mean_length": 1666.19921875, "completions/mean_terminated_length": 699.8482666015625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8230775795852181, "frac_reward_zero_std": 0.65625, "grad_norm": 24.233103509627846, "learning_rate": 1.8362466924127145e-07, "loss": 0.0182, "num_tokens": 2286654038.0, "reward": 0.099609375, "reward_std": 0.12380305677652359, "rewards/accuracy_reward/mean": 0.099609375, "rewards/accuracy_reward/std": 0.29977133870124817, "step": 2411 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1845.0, "completions/mean_length": 1653.6953125, "completions/mean_terminated_length": 563.558837890625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8234189638986089, "frac_reward_zero_std": 0.53125, "grad_norm": 42.865394002176984, "learning_rate": 1.833136488375171e-07, "loss": 0.0271, "num_tokens": 2287570874.0, "reward": 0.06640625, "reward_std": 0.15498074889183044, "rewards/accuracy_reward/mean": 0.06640625, "rewards/accuracy_reward/std": 0.2492343932390213, "step": 2412 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.630859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1966.0, "completions/mean_length": 1557.28515625, "completions/mean_terminated_length": 718.6560668945312, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.8237603482119996, "frac_reward_zero_std": 0.75, "grad_norm": 36.85989035570521, "learning_rate": 1.8300314887797048e-07, "loss": 0.0084, "num_tokens": 2288450668.0, "reward": 0.060546875, "reward_std": 0.09753695130348206, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 2413 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.669921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1981.0, "completions/mean_length": 1591.40234375, "completions/mean_terminated_length": 664.6982421875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8241017325253904, "frac_reward_zero_std": 0.65625, "grad_norm": 49.09549285609956, "learning_rate": 1.8269316980332926e-07, "loss": 0.0093, "num_tokens": 2289348170.0, "reward": 0.052734375, "reward_std": 0.11537161469459534, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 2414 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.658203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1587.0, "completions/mean_length": 1543.24609375, "completions/mean_terminated_length": 571.2342529296875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8244431168387812, "frac_reward_zero_std": 0.78125, "grad_norm": 4.402577956899501, "learning_rate": 1.823837120535523e-07, "loss": 0.004, "num_tokens": 2290215400.0, "reward": 0.072265625, "reward_std": 0.09178133308887482, "rewards/accuracy_reward/mean": 0.072265625, "rewards/accuracy_reward/std": 0.2591804563999176, "step": 2415 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1887.0, "completions/mean_length": 1681.0234375, "completions/mean_terminated_length": 743.1944580078125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8247845011521721, "frac_reward_zero_std": 0.8125, "grad_norm": 9.130866845995918, "learning_rate": 1.820747760678581e-07, "loss": 0.0129, "num_tokens": 2291153140.0, "reward": 0.025390625, "reward_std": 0.06409074366092682, "rewards/accuracy_reward/mean": 0.025390625, "rewards/accuracy_reward/std": 0.15746226906776428, "step": 2416 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.712890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2034.0, "completions/mean_length": 1637.51171875, "completions/mean_terminated_length": 618.2720947265625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8251258854655629, "frac_reward_zero_std": 0.78125, "grad_norm": 19.89982665056698, "learning_rate": 1.8176636228472476e-07, "loss": 0.0181, "num_tokens": 2292069274.0, "reward": 0.064453125, "reward_std": 0.09519492089748383, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 2417 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.634765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1985.0, "completions/mean_length": 1544.58984375, "completions/mean_terminated_length": 669.6791381835938, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.8254672697789537, "frac_reward_zero_std": 0.6875, "grad_norm": 40.93666344695196, "learning_rate": 1.814584711418894e-07, "loss": 0.0016, "num_tokens": 2292936920.0, "reward": 0.091796875, "reward_std": 0.12502044439315796, "rewards/accuracy_reward/mean": 0.091796875, "rewards/accuracy_reward/std": 0.289021372795105, "step": 2418 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.62109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1942.0, "completions/mean_length": 1522.177734375, "completions/mean_terminated_length": 660.2628784179688, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8258086540923445, "frac_reward_zero_std": 0.78125, "grad_norm": 5.5577390552976, "learning_rate": 1.8115110307634695e-07, "loss": 0.0268, "num_tokens": 2293792595.0, "reward": 0.06640625, "reward_std": 0.08493966609239578, "rewards/accuracy_reward/mean": 0.07083333283662796, "rewards/accuracy_reward/std": 0.2568138837814331, "step": 2419 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.634765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1970.0, "completions/mean_length": 1532.4140625, "completions/mean_terminated_length": 636.34228515625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8261500384057353, "frac_reward_zero_std": 0.78125, "grad_norm": 3.73311428328459, "learning_rate": 1.8084425852435044e-07, "loss": 0.0175, "num_tokens": 2294655687.0, "reward": 0.07421875, "reward_std": 0.09094182401895523, "rewards/accuracy_reward/mean": 0.07421875, "rewards/accuracy_reward/std": 0.2623828947544098, "step": 2420 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.701171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1652.044921875, "completions/mean_terminated_length": 722.973876953125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.826491422719126, "frac_reward_zero_std": 0.6875, "grad_norm": 114.67779104481684, "learning_rate": 1.805379379214093e-07, "loss": 0.008, "num_tokens": 2295577390.0, "reward": 0.060546875, "reward_std": 0.10689391195774078, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 2421 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.72265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2016.0, "completions/mean_length": 1668.208984375, "completions/mean_terminated_length": 678.6126708984375, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.8268328070325168, "frac_reward_zero_std": 0.75, "grad_norm": 32.207437152004296, "learning_rate": 1.802321417022899e-07, "loss": 0.0077, "num_tokens": 2296514361.0, "reward": 0.04296875, "reward_std": 0.08655644953250885, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 2422 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.728515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1898.0, "completions/mean_length": 1653.525390625, "completions/mean_terminated_length": 594.9712524414062, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.8271741913459076, "frac_reward_zero_std": 0.8125, "grad_norm": 23.643178987099184, "learning_rate": 1.7992687030101388e-07, "loss": 0.0094, "num_tokens": 2297438278.0, "reward": 0.0625, "reward_std": 0.07528402656316757, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.2422981858253479, "step": 2423 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.70703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2024.0, "completions/mean_length": 1663.80078125, "completions/mean_terminated_length": 736.6000366210938, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8275155756592985, "frac_reward_zero_std": 0.6875, "grad_norm": 11.054214289995855, "learning_rate": 1.7962212415085804e-07, "loss": 0.0314, "num_tokens": 2298369024.0, "reward": 0.04296875, "reward_std": 0.10574321448802948, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 2424 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.666015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2012.0, "completions/mean_length": 1562.69921875, "completions/mean_terminated_length": 594.9356689453125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8278569599726893, "frac_reward_zero_std": 0.71875, "grad_norm": 44.627553102602704, "learning_rate": 1.7931790368435403e-07, "loss": 0.0118, "num_tokens": 2299246086.0, "reward": 0.078125, "reward_std": 0.10787243396043777, "rewards/accuracy_reward/mean": 0.078125, "rewards/accuracy_reward/std": 0.26863065361976624, "step": 2425 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2017.0, "completions/mean_length": 1514.845703125, "completions/mean_terminated_length": 564.4402465820312, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8281983442860801, "frac_reward_zero_std": 0.59375, "grad_norm": 126.73481564879475, "learning_rate": 1.7901420933328696e-07, "loss": 0.0163, "num_tokens": 2300100007.0, "reward": 0.115234375, "reward_std": 0.15972480177879333, "rewards/accuracy_reward/mean": 0.115234375, "rewards/accuracy_reward/std": 0.3196168541908264, "step": 2426 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.662109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1937.0, "completions/mean_length": 1605.845703125, "completions/mean_terminated_length": 739.427734375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8285397285994709, "frac_reward_zero_std": 0.65625, "grad_norm": 85.22665679670817, "learning_rate": 1.787110415286956e-07, "loss": 0.02, "num_tokens": 2300999912.0, "reward": 0.10546875, "reward_std": 0.14591112732887268, "rewards/accuracy_reward/mean": 0.10546875, "rewards/accuracy_reward/std": 0.3074568510055542, "step": 2427 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.65234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1555.646484375, "completions/mean_terminated_length": 631.7921142578125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8288811129128617, "frac_reward_zero_std": 0.71875, "grad_norm": 26.956761066932827, "learning_rate": 1.784084007008711e-07, "loss": 0.0344, "num_tokens": 2301871027.0, "reward": 0.083984375, "reward_std": 0.11674421280622482, "rewards/accuracy_reward/mean": 0.083984375, "rewards/accuracy_reward/std": 0.2776356339454651, "step": 2428 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.666015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2039.0, "completions/mean_length": 1565.625, "completions/mean_terminated_length": 603.6959228515625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8292224972262524, "frac_reward_zero_std": 0.5625, "grad_norm": 12.051030145973305, "learning_rate": 1.781062872793567e-07, "loss": 0.027, "num_tokens": 2302749251.0, "reward": 0.142578125, "reward_std": 0.16356495022773743, "rewards/accuracy_reward/mean": 0.142578125, "rewards/accuracy_reward/std": 0.3499840497970581, "step": 2429 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1922.0, "completions/mean_length": 1599.93359375, "completions/mean_terminated_length": 614.1875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8295638815396432, "frac_reward_zero_std": 0.8125, "grad_norm": 52.219044121701, "learning_rate": 1.778047016929473e-07, "loss": 0.0151, "num_tokens": 2303643105.0, "reward": 0.072265625, "reward_std": 0.0790574848651886, "rewards/accuracy_reward/mean": 0.072265625, "rewards/accuracy_reward/std": 0.2591804563999176, "step": 2430 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.615234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2016.0, "completions/mean_length": 1498.708984375, "completions/mean_terminated_length": 620.4010009765625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.829905265853034, "frac_reward_zero_std": 0.71875, "grad_norm": 21.323735135268663, "learning_rate": 1.7750364436968836e-07, "loss": 0.0009, "num_tokens": 2304488716.0, "reward": 0.068359375, "reward_std": 0.11113345623016357, "rewards/accuracy_reward/mean": 0.068359375, "rewards/accuracy_reward/std": 0.25260838866233826, "step": 2431 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.658203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2013.0, "completions/mean_length": 1559.603515625, "completions/mean_terminated_length": 619.0914306640625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.8302466501664248, "frac_reward_zero_std": 0.65625, "grad_norm": 20.875935578684587, "learning_rate": 1.7720311573687575e-07, "loss": -0.0018, "num_tokens": 2305363937.0, "reward": 0.052734375, "reward_std": 0.1188259869813919, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 2432 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.68359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2031.0, "completions/mean_length": 1596.103515625, "completions/mean_terminated_length": 620.0308837890625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8305880344798157, "frac_reward_zero_std": 0.71875, "grad_norm": 45.548370607970504, "learning_rate": 1.769031162210548e-07, "loss": 0.0028, "num_tokens": 2306261510.0, "reward": 0.041015625, "reward_std": 0.08753519505262375, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 2433 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2000.0, "completions/mean_length": 1634.892578125, "completions/mean_terminated_length": 692.1602783203125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8309294187932065, "frac_reward_zero_std": 0.65625, "grad_norm": 29.62303726552462, "learning_rate": 1.766036462480201e-07, "loss": 0.0408, "num_tokens": 2307179183.0, "reward": 0.0703125, "reward_std": 0.13067752122879028, "rewards/accuracy_reward/mean": 0.0703125, "rewards/accuracy_reward/std": 0.25592297315597534, "step": 2434 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.712890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1963.0, "completions/mean_length": 1654.705078125, "completions/mean_terminated_length": 678.1564331054688, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8312708031065973, "frac_reward_zero_std": 0.625, "grad_norm": 37.45353910406339, "learning_rate": 1.7630470624281442e-07, "loss": -0.0151, "num_tokens": 2308100648.0, "reward": 0.064453125, "reward_std": 0.1277552843093872, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 2435 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1936.0, "completions/mean_length": 1546.240234375, "completions/mean_terminated_length": 620.772216796875, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.8316121874199881, "frac_reward_zero_std": 0.75, "grad_norm": 9.376547185875587, "learning_rate": 1.7600629662972832e-07, "loss": 0.0072, "num_tokens": 2308973827.0, "reward": 0.048828125, "reward_std": 0.08395528793334961, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 2436 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.740234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2026.0, "completions/mean_length": 1684.125, "completions/mean_terminated_length": 647.2180786132812, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8319535717333788, "frac_reward_zero_std": 0.84375, "grad_norm": 29.47996986565397, "learning_rate": 1.757084178322999e-07, "loss": -0.0023, "num_tokens": 2309909251.0, "reward": 0.056640625, "reward_std": 0.06767063587903976, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 2437 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.73046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2032.0, "completions/mean_length": 1657.455078125, "completions/mean_terminated_length": 599.021728515625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.8322949560467696, "frac_reward_zero_std": 0.71875, "grad_norm": 16.292955670914946, "learning_rate": 1.754110702733134e-07, "loss": 0.0115, "num_tokens": 2310840076.0, "reward": 0.05078125, "reward_std": 0.10295412689447403, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 2438 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.763671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1829.0, "completions/mean_length": 1715.990234375, "completions/mean_terminated_length": 643.1322021484375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8326363403601604, "frac_reward_zero_std": 0.6875, "grad_norm": 84.10901240630035, "learning_rate": 1.7511425437479946e-07, "loss": 0.0158, "num_tokens": 2311800247.0, "reward": 0.0703125, "reward_std": 0.11305683851242065, "rewards/accuracy_reward/mean": 0.0703125, "rewards/accuracy_reward/std": 0.25592297315597534, "step": 2439 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.755859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1962.0, "completions/mean_length": 1703.0625, "completions/mean_terminated_length": 635.1360473632812, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8329777246735512, "frac_reward_zero_std": 0.78125, "grad_norm": 28.049883394507432, "learning_rate": 1.7481797055803382e-07, "loss": 0.0184, "num_tokens": 2312747815.0, "reward": 0.046875, "reward_std": 0.07966843992471695, "rewards/accuracy_reward/mean": 0.046875, "rewards/accuracy_reward/std": 0.21157780289649963, "step": 2440 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.65625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1990.0, "completions/mean_length": 1544.76171875, "completions/mean_terminated_length": 584.0341186523438, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.8333191089869421, "frac_reward_zero_std": 0.5625, "grad_norm": 14.721422240156052, "learning_rate": 1.7452221924353733e-07, "loss": 0.0318, "num_tokens": 2313620301.0, "reward": 0.126953125, "reward_std": 0.17244866490364075, "rewards/accuracy_reward/mean": 0.13104838132858276, "rewards/accuracy_reward/std": 0.3377939462661743, "step": 2441 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.73828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2021.0, "completions/mean_length": 1706.8984375, "completions/mean_terminated_length": 744.6865234375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8336604933003329, "frac_reward_zero_std": 0.875, "grad_norm": 7.546618130082427, "learning_rate": 1.7422700085107485e-07, "loss": 0.0059, "num_tokens": 2314570681.0, "reward": 0.048828125, "reward_std": 0.043847277760505676, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 2442 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.73046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1829.0, "completions/mean_length": 1680.009765625, "completions/mean_terminated_length": 683.1087036132812, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8340018776137237, "frac_reward_zero_std": 0.78125, "grad_norm": 24.690283639551385, "learning_rate": 1.7393231579965467e-07, "loss": -0.0042, "num_tokens": 2315520990.0, "reward": 0.0234375, "reward_std": 0.06805649399757385, "rewards/accuracy_reward/mean": 0.0234375, "rewards/accuracy_reward/std": 0.15143637359142303, "step": 2443 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.666015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2037.0, "completions/mean_length": 1585.6171875, "completions/mean_terminated_length": 663.5555419921875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8343432619271145, "frac_reward_zero_std": 0.78125, "grad_norm": 57.897965728013894, "learning_rate": 1.736381645075286e-07, "loss": 0.0139, "num_tokens": 2316409178.0, "reward": 0.044921875, "reward_std": 0.08780106902122498, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 2444 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.75390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2040.0, "completions/mean_length": 1700.16796875, "completions/mean_terminated_length": 634.5873413085938, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.8346846462405052, "frac_reward_zero_std": 0.6875, "grad_norm": 15.03079739787849, "learning_rate": 1.733445473921904e-07, "loss": 0.0248, "num_tokens": 2317353712.0, "reward": 0.056640625, "reward_std": 0.10837717354297638, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 2445 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.759765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1870.0, "completions/mean_length": 1707.330078125, "completions/mean_terminated_length": 629.9268188476562, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.835026030553896, "frac_reward_zero_std": 0.78125, "grad_norm": 7.0236254032041066, "learning_rate": 1.7305146487037603e-07, "loss": 0.0237, "num_tokens": 2318301001.0, "reward": 0.044921875, "reward_std": 0.08164606243371964, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 2446 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1864.0, "completions/mean_length": 1691.9453125, "completions/mean_terminated_length": 666.9393920898438, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.8353674148672868, "frac_reward_zero_std": 0.75, "grad_norm": 114.86323201527865, "learning_rate": 1.727589173580625e-07, "loss": 0.0156, "num_tokens": 2319242669.0, "reward": 0.037109375, "reward_std": 0.08295752108097076, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 2447 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.68359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2039.0, "completions/mean_length": 1602.259765625, "completions/mean_terminated_length": 639.24072265625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8357087991806776, "frac_reward_zero_std": 0.625, "grad_norm": 12.385882218391371, "learning_rate": 1.724669052704673e-07, "loss": 0.0362, "num_tokens": 2320133314.0, "reward": 0.08203125, "reward_std": 0.13352441787719727, "rewards/accuracy_reward/mean": 0.08203125, "rewards/accuracy_reward/std": 0.2746807038784027, "step": 2448 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.771484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1970.0, "completions/mean_length": 1711.5546875, "completions/mean_terminated_length": 575.6923217773438, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8360501834940685, "frac_reward_zero_std": 0.84375, "grad_norm": 1.4256683652082736, "learning_rate": 1.7217542902204847e-07, "loss": 0.0229, "num_tokens": 2321092654.0, "reward": 0.05078125, "reward_std": 0.07383356243371964, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 2449 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.73046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1951.0, "completions/mean_length": 1675.33984375, "completions/mean_terminated_length": 665.3768310546875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.8363915678074593, "frac_reward_zero_std": 0.84375, "grad_norm": 10.29077629259919, "learning_rate": 1.7188448902650287e-07, "loss": 0.012, "num_tokens": 2322026860.0, "reward": 0.01171875, "reward_std": 0.04192390665411949, "rewards/accuracy_reward/mean": 0.01171875, "rewards/accuracy_reward/std": 0.10772226005792618, "step": 2450 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.73046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1825.0, "completions/mean_length": 1671.484375, "completions/mean_terminated_length": 651.0724487304688, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8367329521208501, "frac_reward_zero_std": 0.6875, "grad_norm": 31.571270963268077, "learning_rate": 1.7159408569676704e-07, "loss": 0.0162, "num_tokens": 2322953828.0, "reward": 0.064453125, "reward_std": 0.1156046986579895, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 2451 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.72265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2031.0, "completions/mean_length": 1681.0625, "completions/mean_terminated_length": 724.9577026367188, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.8370743364342409, "frac_reward_zero_std": 0.75, "grad_norm": 54.7029599860654, "learning_rate": 1.71304219445015e-07, "loss": 0.0053, "num_tokens": 2323898308.0, "reward": 0.05078125, "reward_std": 0.0961260050535202, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 2452 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.677734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1935.0, "completions/mean_length": 1593.720703125, "completions/mean_terminated_length": 638.3575439453125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8374157207476316, "frac_reward_zero_std": 0.71875, "grad_norm": 40.60600258790086, "learning_rate": 1.7101489068265935e-07, "loss": 0.0006, "num_tokens": 2324791125.0, "reward": 0.056640625, "reward_std": 0.08318261057138443, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 2453 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1929.0, "completions/mean_length": 1652.125, "completions/mean_terminated_length": 678.4865112304688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8377571050610224, "frac_reward_zero_std": 0.71875, "grad_norm": 51.89640687927489, "learning_rate": 1.7072609982034874e-07, "loss": 0.0269, "num_tokens": 2325711189.0, "reward": 0.072265625, "reward_std": 0.11046035587787628, "rewards/accuracy_reward/mean": 0.072265625, "rewards/accuracy_reward/std": 0.2591804563999176, "step": 2454 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.76953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1999.0, "completions/mean_length": 1726.75390625, "completions/mean_terminated_length": 654.11865234375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8380984893744132, "frac_reward_zero_std": 0.8125, "grad_norm": 22.367061206155213, "learning_rate": 1.7043784726796934e-07, "loss": 0.0158, "num_tokens": 2326676951.0, "reward": 0.041015625, "reward_std": 0.07322713732719421, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 2455 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.689453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1904.0, "completions/mean_length": 1581.43359375, "completions/mean_terminated_length": 545.5974731445312, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.838439873687804, "frac_reward_zero_std": 0.59375, "grad_norm": 98.02973095646546, "learning_rate": 1.7015013343464302e-07, "loss": -0.0102, "num_tokens": 2327561781.0, "reward": 0.060546875, "reward_std": 0.13550205528736115, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.24230584502220154, "step": 2456 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.69140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2004.0, "completions/mean_length": 1597.53125, "completions/mean_terminated_length": 588.253173828125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8387812580011949, "frac_reward_zero_std": 0.65625, "grad_norm": 34.01524040146943, "learning_rate": 1.698629587287266e-07, "loss": 0.043, "num_tokens": 2328449317.0, "reward": 0.099609375, "reward_std": 0.1253339648246765, "rewards/accuracy_reward/mean": 0.099609375, "rewards/accuracy_reward/std": 0.29977133870124817, "step": 2457 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.701171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1964.0, "completions/mean_length": 1618.376953125, "completions/mean_terminated_length": 610.6732177734375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8391226423145857, "frac_reward_zero_std": 0.75, "grad_norm": 23.28112767502058, "learning_rate": 1.6957632355781243e-07, "loss": 0.0167, "num_tokens": 2329355526.0, "reward": 0.060546875, "reward_std": 0.09759023040533066, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 2458 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.69921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1898.0, "completions/mean_length": 1624.89453125, "completions/mean_terminated_length": 641.3117065429688, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8394640266279765, "frac_reward_zero_std": 0.71875, "grad_norm": 7.013998671697308, "learning_rate": 1.6929022832872653e-07, "loss": 0.0125, "num_tokens": 2330265648.0, "reward": 0.07421875, "reward_std": 0.08549975603818893, "rewards/accuracy_reward/mean": 0.07661290466785431, "rewards/accuracy_reward/std": 0.2662447690963745, "step": 2459 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2035.0, "completions/mean_length": 1567.09765625, "completions/mean_terminated_length": 680.1000366210938, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.8398054109413673, "frac_reward_zero_std": 0.71875, "grad_norm": 14.973416587170767, "learning_rate": 1.6900467344752872e-07, "loss": 0.0095, "num_tokens": 2331139458.0, "reward": 0.0390625, "reward_std": 0.09039659798145294, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 2460 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.736328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1803.0, "completions/mean_length": 1666.265625, "completions/mean_terminated_length": 600.2369995117188, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.840146795254758, "frac_reward_zero_std": 0.71875, "grad_norm": 6.299233016236517, "learning_rate": 1.6871965931951178e-07, "loss": 0.0255, "num_tokens": 2332060938.0, "reward": 0.0546875, "reward_std": 0.0993933230638504, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 2461 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1968.0, "completions/mean_length": 1616.279296875, "completions/mean_terminated_length": 593.7828979492188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8404881795681488, "frac_reward_zero_std": 0.9375, "grad_norm": 2.8045307953462024, "learning_rate": 1.684351863492014e-07, "loss": 0.0025, "num_tokens": 2332962521.0, "reward": 0.021484375, "reward_std": 0.0234375, "rewards/accuracy_reward/mean": 0.02217741869390011, "rewards/accuracy_reward/std": 0.14740893244743347, "step": 2462 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.708984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1679.0, "completions/mean_length": 1624.1640625, "completions/mean_terminated_length": 591.5972900390625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8408295638815396, "frac_reward_zero_std": 0.78125, "grad_norm": 42.726586959532604, "learning_rate": 1.6815125494035494e-07, "loss": 0.0067, "num_tokens": 2333875181.0, "reward": 0.041015625, "reward_std": 0.08521346747875214, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 2463 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.712890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1806.0, "completions/mean_length": 1627.390625, "completions/mean_terminated_length": 583.0203857421875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8411709481949304, "frac_reward_zero_std": 0.8125, "grad_norm": 24.174258047440986, "learning_rate": 1.678678654959609e-07, "loss": 0.003, "num_tokens": 2334783669.0, "reward": 0.04296875, "reward_std": 0.06398563086986542, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 2464 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.802734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1987.0, "completions/mean_length": 1762.734375, "completions/mean_terminated_length": 601.9010009765625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8415123325083212, "frac_reward_zero_std": 0.78125, "grad_norm": 23.78048564079735, "learning_rate": 1.6758501841823902e-07, "loss": -0.0017, "num_tokens": 2335775469.0, "reward": 0.029296875, "reward_std": 0.07108421623706818, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 2465 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2007.0, "completions/mean_length": 1666.48046875, "completions/mean_terminated_length": 691.4861450195312, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8418537168217121, "frac_reward_zero_std": 0.75, "grad_norm": 24.300035680049852, "learning_rate": 1.6730271410863864e-07, "loss": 0.009, "num_tokens": 2336708163.0, "reward": 0.0390625, "reward_std": 0.07889671623706818, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 2466 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.70703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1896.0, "completions/mean_length": 1634.92578125, "completions/mean_terminated_length": 638.0400390625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.8421951011351029, "frac_reward_zero_std": 0.78125, "grad_norm": 99.19390883899514, "learning_rate": 1.6702095296783942e-07, "loss": 0.0093, "num_tokens": 2337620109.0, "reward": 0.060546875, "reward_std": 0.07669496536254883, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 2467 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1784.0, "completions/mean_length": 1722.70703125, "completions/mean_terminated_length": 704.8547973632812, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.8425364854484937, "frac_reward_zero_std": 0.71875, "grad_norm": 47.42924908200362, "learning_rate": 1.6673973539574953e-07, "loss": 0.0133, "num_tokens": 2338585607.0, "reward": 0.044921875, "reward_std": 0.09935424476861954, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 2468 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.69140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1700.0, "completions/mean_length": 1585.93359375, "completions/mean_terminated_length": 550.6708984375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8428778697618844, "frac_reward_zero_std": 0.6875, "grad_norm": 13.8061170052062, "learning_rate": 1.6645906179150592e-07, "loss": 0.0202, "num_tokens": 2339475109.0, "reward": 0.052734375, "reward_std": 0.11073969304561615, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 2469 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.712890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1708.0, "completions/mean_length": 1641.55859375, "completions/mean_terminated_length": 632.3673095703125, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.8432192540752752, "frac_reward_zero_std": 0.65625, "grad_norm": 31.773593954625255, "learning_rate": 1.661789325534737e-07, "loss": 0.0372, "num_tokens": 2340385523.0, "reward": 0.1015625, "reward_std": 0.14435848593711853, "rewards/accuracy_reward/mean": 0.10483870655298233, "rewards/accuracy_reward/std": 0.30665475130081177, "step": 2470 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.76171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2019.0, "completions/mean_length": 1733.9140625, "completions/mean_terminated_length": 729.8688354492188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.843560638388666, "frac_reward_zero_std": 0.8125, "grad_norm": 27.74800951683062, "learning_rate": 1.6589934807924482e-07, "loss": 0.0095, "num_tokens": 2341348071.0, "reward": 0.03515625, "reward_std": 0.07135801017284393, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 2471 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.744140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1622.0, "completions/mean_length": 1669.490234375, "completions/mean_terminated_length": 568.6336059570312, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8439020227020568, "frac_reward_zero_std": 0.71875, "grad_norm": 69.43353361004944, "learning_rate": 1.6562030876563843e-07, "loss": 0.0171, "num_tokens": 2342277586.0, "reward": 0.06640625, "reward_std": 0.10821876674890518, "rewards/accuracy_reward/mean": 0.06640625, "rewards/accuracy_reward/std": 0.2492343932390213, "step": 2472 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1819.0, "completions/mean_length": 1729.23828125, "completions/mean_terminated_length": 641.0516967773438, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.8442434070154476, "frac_reward_zero_std": 0.71875, "grad_norm": 108.39637458379629, "learning_rate": 1.653418150086996e-07, "loss": -0.0024, "num_tokens": 2343238652.0, "reward": 0.0625, "reward_std": 0.09908141195774078, "rewards/accuracy_reward/mean": 0.06451612710952759, "rewards/accuracy_reward/std": 0.2459181249141693, "step": 2473 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.65234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1952.0, "completions/mean_length": 1549.693359375, "completions/mean_terminated_length": 614.6685180664062, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8445847913288385, "frac_reward_zero_std": 0.59375, "grad_norm": 58.10240198511226, "learning_rate": 1.6506386720369953e-07, "loss": 0.0369, "num_tokens": 2344104735.0, "reward": 0.10546875, "reward_std": 0.1544739454984665, "rewards/accuracy_reward/mean": 0.10546875, "rewards/accuracy_reward/std": 0.3074568510055542, "step": 2474 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.669921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2033.0, "completions/mean_length": 1586.568359375, "completions/mean_terminated_length": 650.0532836914062, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8449261756422293, "frac_reward_zero_std": 0.8125, "grad_norm": 41.46929513024965, "learning_rate": 1.6478646574513409e-07, "loss": -0.0125, "num_tokens": 2345004674.0, "reward": 0.041015625, "reward_std": 0.062062256038188934, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 2475 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.78515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1949.0, "completions/mean_length": 1738.802734375, "completions/mean_terminated_length": 608.8272705078125, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.8452675599556201, "frac_reward_zero_std": 0.625, "grad_norm": 82.17590699252654, "learning_rate": 1.6450961102672394e-07, "loss": 0.0302, "num_tokens": 2345981037.0, "reward": 0.0859375, "reward_std": 0.1245543360710144, "rewards/accuracy_reward/mean": 0.0859375, "rewards/accuracy_reward/std": 0.28054583072662354, "step": 2476 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.76953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1863.0, "completions/mean_length": 1718.568359375, "completions/mean_terminated_length": 618.6016845703125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8456089442690108, "frac_reward_zero_std": 0.78125, "grad_norm": 38.68191746576925, "learning_rate": 1.6423330344141401e-07, "loss": 0.013, "num_tokens": 2346938416.0, "reward": 0.041015625, "reward_std": 0.06849660724401474, "rewards/accuracy_reward/mean": 0.04233871027827263, "rewards/accuracy_reward/std": 0.2015640139579773, "step": 2477 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.736328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1973.0, "completions/mean_length": 1696.4140625, "completions/mean_terminated_length": 714.5777587890625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8459503285824016, "frac_reward_zero_std": 0.75, "grad_norm": 116.31613822570915, "learning_rate": 1.6395754338137203e-07, "loss": 0.0031, "num_tokens": 2347885860.0, "reward": 0.0390625, "reward_std": 0.08917921781539917, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 2478 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.697265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1793.0, "completions/mean_length": 1590.57421875, "completions/mean_terminated_length": 537.0193481445312, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8462917128957924, "frac_reward_zero_std": 0.59375, "grad_norm": 69.25798212378204, "learning_rate": 1.6368233123798913e-07, "loss": 0.0213, "num_tokens": 2348766746.0, "reward": 0.068359375, "reward_std": 0.14085254073143005, "rewards/accuracy_reward/mean": 0.068359375, "rewards/accuracy_reward/std": 0.25260838866233826, "step": 2479 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.603515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2018.0, "completions/mean_length": 1501.287109375, "completions/mean_terminated_length": 669.0985107421875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.8466330972091832, "frac_reward_zero_std": 0.75, "grad_norm": 27.80232104779852, "learning_rate": 1.634076674018785e-07, "loss": 0.0246, "num_tokens": 2349610973.0, "reward": 0.056640625, "reward_std": 0.09331226348876953, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 2480 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1896.0, "completions/mean_length": 1622.630859375, "completions/mean_terminated_length": 535.576416015625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.846974481522574, "frac_reward_zero_std": 0.875, "grad_norm": 117.39871145366247, "learning_rate": 1.6313355226287535e-07, "loss": 0.0168, "num_tokens": 2350519328.0, "reward": 0.048828125, "reward_std": 0.05463561788201332, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 2481 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.779296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1913.0, "completions/mean_length": 1717.986328125, "completions/mean_terminated_length": 552.716796875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8473158658359649, "frac_reward_zero_std": 0.875, "grad_norm": 7.967204647278244, "learning_rate": 1.6285998621003581e-07, "loss": 0.0037, "num_tokens": 2351478089.0, "reward": 0.009765625, "reward_std": 0.03411140665411949, "rewards/accuracy_reward/mean": 0.009765625, "rewards/accuracy_reward/std": 0.09843364357948303, "step": 2482 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.763671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1955.0, "completions/mean_length": 1734.9765625, "completions/mean_terminated_length": 723.4710693359375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8476572501493557, "frac_reward_zero_std": 0.78125, "grad_norm": 13.644900709205602, "learning_rate": 1.6258696963163704e-07, "loss": 0.0205, "num_tokens": 2352449341.0, "reward": 0.0625, "reward_std": 0.09110813587903976, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.2422981858253479, "step": 2483 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.720703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1894.0, "completions/mean_length": 1675.392578125, "completions/mean_terminated_length": 713.9090576171875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.8479986344627465, "frac_reward_zero_std": 0.6875, "grad_norm": 313.4522522875926, "learning_rate": 1.6231450291517617e-07, "loss": 0.0039, "num_tokens": 2353386534.0, "reward": 0.072265625, "reward_std": 0.10695268213748932, "rewards/accuracy_reward/mean": 0.072265625, "rewards/accuracy_reward/std": 0.2591804563999176, "step": 2484 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1911.0, "completions/mean_length": 1575.779296875, "completions/mean_terminated_length": 608.8511962890625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8483400187761372, "frac_reward_zero_std": 0.71875, "grad_norm": 56.12616883071651, "learning_rate": 1.6204258644736966e-07, "loss": 0.0162, "num_tokens": 2354271285.0, "reward": 0.060546875, "reward_std": 0.09633205831050873, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 2485 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1911.0, "completions/mean_length": 1639.482421875, "completions/mean_terminated_length": 615.390380859375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.848681403089528, "frac_reward_zero_std": 0.71875, "grad_norm": 3.873614259818, "learning_rate": 1.6177122061415337e-07, "loss": 0.0327, "num_tokens": 2355181916.0, "reward": 0.078125, "reward_std": 0.11195248365402222, "rewards/accuracy_reward/mean": 0.08064515888690948, "rewards/accuracy_reward/std": 0.2725643217563629, "step": 2486 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.75390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2021.0, "completions/mean_length": 1720.359375, "completions/mean_terminated_length": 716.6349487304688, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.8490227874029188, "frac_reward_zero_std": 0.78125, "grad_norm": 36.28393802281751, "learning_rate": 1.615004058006812e-07, "loss": 0.0171, "num_tokens": 2356148356.0, "reward": 0.048828125, "reward_std": 0.08808182179927826, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 2487 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.67578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2014.0, "completions/mean_length": 1581.90625, "completions/mean_terminated_length": 610.4096069335938, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8493641717163096, "frac_reward_zero_std": 0.875, "grad_norm": 2.365506156421668, "learning_rate": 1.6123014239132568e-07, "loss": 0.0021, "num_tokens": 2357029412.0, "reward": 0.02734375, "reward_std": 0.04847269132733345, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 2488 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.73828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2026.0, "completions/mean_length": 1681.94921875, "completions/mean_terminated_length": 649.3582153320312, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8497055560297004, "frac_reward_zero_std": 0.59375, "grad_norm": 126.78743633048502, "learning_rate": 1.6096043076967592e-07, "loss": 0.0075, "num_tokens": 2357968330.0, "reward": 0.099609375, "reward_std": 0.14469268918037415, "rewards/accuracy_reward/mean": 0.099609375, "rewards/accuracy_reward/std": 0.29977133870124817, "step": 2489 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2019.0, "completions/mean_length": 1739.74609375, "completions/mean_terminated_length": 732.7833862304688, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.8500469403430913, "frac_reward_zero_std": 0.875, "grad_norm": 3.1716992311816425, "learning_rate": 1.6069127131853846e-07, "loss": 0.0182, "num_tokens": 2358930376.0, "reward": 0.025390625, "reward_std": 0.04957009106874466, "rewards/accuracy_reward/mean": 0.025390625, "rewards/accuracy_reward/std": 0.15746226906776428, "step": 2490 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.75, "completions/max_length": 2048.0, "completions/max_terminated_length": 1891.0, "completions/mean_length": 1692.513671875, "completions/mean_terminated_length": 626.5234375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8503883246564821, "frac_reward_zero_std": 0.75, "grad_norm": 84.24221189164871, "learning_rate": 1.6042266441993583e-07, "loss": 0.0129, "num_tokens": 2359877935.0, "reward": 0.05859375, "reward_std": 0.10216332226991653, "rewards/accuracy_reward/mean": 0.05859375, "rewards/accuracy_reward/std": 0.23509246110916138, "step": 2491 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1894.0, "completions/mean_length": 1595.748046875, "completions/mean_terminated_length": 636.0914306640625, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.8507297089698729, "frac_reward_zero_std": 0.625, "grad_norm": 198.28256082638302, "learning_rate": 1.6015461045510626e-07, "loss": 0.0196, "num_tokens": 2360779406.0, "reward": 0.103515625, "reward_std": 0.14421284198760986, "rewards/accuracy_reward/mean": 0.103515625, "rewards/accuracy_reward/std": 0.30492907762527466, "step": 2492 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.72265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1941.0, "completions/mean_length": 1626.03125, "completions/mean_terminated_length": 526.5352172851562, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8510710932832636, "frac_reward_zero_std": 0.8125, "grad_norm": 10.510558003897218, "learning_rate": 1.598871098045036e-07, "loss": 0.0047, "num_tokens": 2361686302.0, "reward": 0.013671875, "reward_std": 0.04973640665411949, "rewards/accuracy_reward/mean": 0.013671875, "rewards/accuracy_reward/std": 0.1162383034825325, "step": 2493 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.759765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1760.0, "completions/mean_length": 1689.2265625, "completions/mean_terminated_length": 554.569091796875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.8514124775966544, "frac_reward_zero_std": 0.71875, "grad_norm": 18.77744092290952, "learning_rate": 1.5962016284779578e-07, "loss": 0.0092, "num_tokens": 2362627874.0, "reward": 0.091796875, "reward_std": 0.1260877102613449, "rewards/accuracy_reward/mean": 0.0947580635547638, "rewards/accuracy_reward/std": 0.29317617416381836, "step": 2494 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.62890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1999.0, "completions/mean_length": 1507.29296875, "completions/mean_terminated_length": 590.9368896484375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8517538619100452, "frac_reward_zero_std": 0.6875, "grad_norm": 26.690213498562148, "learning_rate": 1.5935376996386552e-07, "loss": 0.0189, "num_tokens": 2363471672.0, "reward": 0.064453125, "reward_std": 0.10794496536254883, "rewards/accuracy_reward/mean": 0.06653226166963577, "rewards/accuracy_reward/std": 0.2494617998600006, "step": 2495 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.75, "completions/max_length": 2048.0, "completions/max_terminated_length": 1898.0, "completions/mean_length": 1699.416015625, "completions/mean_terminated_length": 653.6640625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.852095246223436, "frac_reward_zero_std": 0.75, "grad_norm": 23.369315616820842, "learning_rate": 1.590879315308086e-07, "loss": 0.0159, "num_tokens": 2364424605.0, "reward": 0.080078125, "reward_std": 0.10760685801506042, "rewards/accuracy_reward/mean": 0.080078125, "rewards/accuracy_reward/std": 0.271679550409317, "step": 2496 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.693359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1972.0, "completions/mean_length": 1607.916015625, "completions/mean_terminated_length": 612.8216552734375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8524366305368268, "frac_reward_zero_std": 0.84375, "grad_norm": 25.250005551357134, "learning_rate": 1.5882264792593397e-07, "loss": 0.0097, "num_tokens": 2365324274.0, "reward": 0.0390625, "reward_std": 0.06355905532836914, "rewards/accuracy_reward/mean": 0.04032257944345474, "rewards/accuracy_reward/std": 0.19691328704357147, "step": 2497 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.748046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1930.0, "completions/mean_length": 1687.103515625, "completions/mean_terminated_length": 615.9767456054688, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.8527780148502176, "frac_reward_zero_std": 0.8125, "grad_norm": 58.161082483966034, "learning_rate": 1.5855791952576342e-07, "loss": 0.0048, "num_tokens": 2366277175.0, "reward": 0.044921875, "reward_std": 0.06260748207569122, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 2498 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.73046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1918.0, "completions/mean_length": 1684.2890625, "completions/mean_terminated_length": 698.5797119140625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.8531193991636085, "frac_reward_zero_std": 0.84375, "grad_norm": 37.843486555756215, "learning_rate": 1.582937467060302e-07, "loss": 0.0031, "num_tokens": 2367221611.0, "reward": 0.064453125, "reward_std": 0.060957904905080795, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 2499 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.787109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1881.0, "completions/mean_length": 1754.24609375, "completions/mean_terminated_length": 668.6972045898438, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.8534607834769993, "frac_reward_zero_std": 0.84375, "grad_norm": 45.53725030401169, "learning_rate": 1.5803012984167963e-07, "loss": 0.0124, "num_tokens": 2368205209.0, "reward": 0.05078125, "reward_std": 0.062054343521595, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 2500 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1966.0, "completions/mean_length": 1637.3359375, "completions/mean_terminated_length": 627.3243408203125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.85380216779039, "frac_reward_zero_std": 0.78125, "grad_norm": 23.381935556505667, "learning_rate": 1.5776706930686738e-07, "loss": -0.0021, "num_tokens": 2369120229.0, "reward": 0.048828125, "reward_std": 0.08042772114276886, "rewards/accuracy_reward/mean": 0.05040322616696358, "rewards/accuracy_reward/std": 0.21899642050266266, "step": 2501 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.681640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1816.0, "completions/mean_length": 1587.3828125, "completions/mean_terminated_length": 601.1533813476562, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8541435521037808, "frac_reward_zero_std": 0.84375, "grad_norm": 102.70727538071121, "learning_rate": 1.5750456547495995e-07, "loss": -0.0049, "num_tokens": 2370012601.0, "reward": 0.041015625, "reward_std": 0.06642046570777893, "rewards/accuracy_reward/mean": 0.04233871027827263, "rewards/accuracy_reward/std": 0.2015640139579773, "step": 2502 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.57421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1945.0, "completions/mean_length": 1442.87890625, "completions/mean_terminated_length": 626.7981567382812, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.8544849364171716, "frac_reward_zero_std": 0.6875, "grad_norm": 66.00511986661476, "learning_rate": 1.572426187185334e-07, "loss": 0.012, "num_tokens": 2370823515.0, "reward": 0.0859375, "reward_std": 0.12740197777748108, "rewards/accuracy_reward/mean": 0.08870967477560043, "rewards/accuracy_reward/std": 0.2846112847328186, "step": 2503 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.65234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1896.0, "completions/mean_length": 1560.91796875, "completions/mean_terminated_length": 646.955078125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8548263207305624, "frac_reward_zero_std": 0.9375, "grad_norm": 3.9006223281016066, "learning_rate": 1.5698122940937325e-07, "loss": 0.005, "num_tokens": 2371692049.0, "reward": 0.013671875, "reward_std": 0.0234375, "rewards/accuracy_reward/mean": 0.013671875, "rewards/accuracy_reward/std": 0.1162383034825325, "step": 2504 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1948.0, "completions/mean_length": 1723.134765625, "completions/mean_terminated_length": 661.9083862304688, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8551677050439532, "frac_reward_zero_std": 0.75, "grad_norm": 29.66417409840493, "learning_rate": 1.5672039791847385e-07, "loss": 0.0113, "num_tokens": 2372644150.0, "reward": 0.052734375, "reward_std": 0.08868034183979034, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 2505 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.69140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1976.0, "completions/mean_length": 1617.95703125, "completions/mean_terminated_length": 654.4430541992188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.855509089357344, "frac_reward_zero_std": 0.84375, "grad_norm": 69.17180149536003, "learning_rate": 1.5646012461603773e-07, "loss": 0.0094, "num_tokens": 2373550160.0, "reward": 0.017578125, "reward_std": 0.050508126616477966, "rewards/accuracy_reward/mean": 0.017578125, "rewards/accuracy_reward/std": 0.13154059648513794, "step": 2506 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.728515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1979.0, "completions/mean_length": 1668.06640625, "completions/mean_terminated_length": 648.5324096679688, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.8558504736707349, "frac_reward_zero_std": 0.8125, "grad_norm": 14.844015147095856, "learning_rate": 1.5620040987147536e-07, "loss": 0.0215, "num_tokens": 2374483346.0, "reward": 0.03125, "reward_std": 0.06657323986291885, "rewards/accuracy_reward/mean": 0.032258063554763794, "rewards/accuracy_reward/std": 0.17686308920383453, "step": 2507 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.693359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1977.0, "completions/mean_length": 1606.63671875, "completions/mean_terminated_length": 608.6497192382812, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8561918579841257, "frac_reward_zero_std": 0.71875, "grad_norm": 23.364996652415677, "learning_rate": 1.559412540534043e-07, "loss": 0.0005, "num_tokens": 2375383816.0, "reward": 0.0546875, "reward_std": 0.09908141195774078, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 2508 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.75, "completions/max_length": 2048.0, "completions/max_terminated_length": 1953.0, "completions/mean_length": 1696.10546875, "completions/mean_terminated_length": 640.421875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8565332422975165, "frac_reward_zero_std": 0.84375, "grad_norm": 17.16084546199387, "learning_rate": 1.5568265752964865e-07, "loss": 0.0067, "num_tokens": 2376329614.0, "reward": 0.029296875, "reward_std": 0.05870648846030235, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 2509 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1472.0, "completions/mean_length": 1746.1953125, "completions/mean_terminated_length": 562.75, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.8568746266109072, "frac_reward_zero_std": 0.71875, "grad_norm": 29.371474981579752, "learning_rate": 1.5542462066723912e-07, "loss": 0.0284, "num_tokens": 2377307122.0, "reward": 0.09375, "reward_std": 0.1272030919790268, "rewards/accuracy_reward/mean": 0.09375, "rewards/accuracy_reward/std": 0.29176566004753113, "step": 2510 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1872.0, "completions/mean_length": 1629.189453125, "completions/mean_terminated_length": 599.1419067382812, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.857216010924298, "frac_reward_zero_std": 0.8125, "grad_norm": 14.070377713547375, "learning_rate": 1.551671438324116e-07, "loss": 0.0028, "num_tokens": 2378229907.0, "reward": 0.052734375, "reward_std": 0.06271954625844955, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 2511 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.728515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1882.0, "completions/mean_length": 1686.822265625, "completions/mean_terminated_length": 717.6187133789062, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8575573952376888, "frac_reward_zero_std": 0.71875, "grad_norm": 47.8749366852237, "learning_rate": 1.549102273906076e-07, "loss": 0.0095, "num_tokens": 2379166776.0, "reward": 0.05078125, "reward_std": 0.10232867300510406, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 2512 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.755859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1959.0, "completions/mean_length": 1691.681640625, "completions/mean_terminated_length": 588.52001953125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8578987795510796, "frac_reward_zero_std": 0.75, "grad_norm": 31.53900131626096, "learning_rate": 1.5465387170647284e-07, "loss": 0.0096, "num_tokens": 2380113589.0, "reward": 0.05859375, "reward_std": 0.10085200518369675, "rewards/accuracy_reward/mean": 0.05859375, "rewards/accuracy_reward/std": 0.23509246110916138, "step": 2513 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.751953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1986.0, "completions/mean_length": 1708.517578125, "completions/mean_terminated_length": 679.3779296875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.8582401638644704, "frac_reward_zero_std": 0.8125, "grad_norm": 83.50638973783518, "learning_rate": 1.5439807714385747e-07, "loss": 0.0194, "num_tokens": 2381069694.0, "reward": 0.0625, "reward_std": 0.07863742858171463, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.2422981858253479, "step": 2514 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2021.0, "completions/mean_length": 1691.732421875, "completions/mean_terminated_length": 706.7573852539062, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8585815481778613, "frac_reward_zero_std": 0.8125, "grad_norm": 71.44569889232295, "learning_rate": 1.54142844065815e-07, "loss": 0.0243, "num_tokens": 2382012053.0, "reward": 0.048828125, "reward_std": 0.06519509106874466, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 2515 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.744140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1959.0, "completions/mean_length": 1688.953125, "completions/mean_terminated_length": 644.7022705078125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8589229324912521, "frac_reward_zero_std": 0.875, "grad_norm": 20.395569760908234, "learning_rate": 1.5388817283460205e-07, "loss": 0.0098, "num_tokens": 2382955405.0, "reward": 0.017578125, "reward_std": 0.043680962175130844, "rewards/accuracy_reward/mean": 0.017578125, "rewards/accuracy_reward/std": 0.13154059648513794, "step": 2516 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.794921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1680.0, "completions/mean_length": 1748.953125, "completions/mean_terminated_length": 589.79052734375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8592643168046429, "frac_reward_zero_std": 0.8125, "grad_norm": 24.08760944224268, "learning_rate": 1.5363406381167798e-07, "loss": -0.0025, "num_tokens": 2383926293.0, "reward": 0.046875, "reward_std": 0.06948098540306091, "rewards/accuracy_reward/mean": 0.046875, "rewards/accuracy_reward/std": 0.21157780289649963, "step": 2517 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.759765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1978.0, "completions/mean_length": 1709.533203125, "completions/mean_terminated_length": 639.0975341796875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8596057011180336, "frac_reward_zero_std": 0.6875, "grad_norm": 53.0481258390873, "learning_rate": 1.533805173577039e-07, "loss": 0.0253, "num_tokens": 2384881878.0, "reward": 0.09765625, "reward_std": 0.12600576877593994, "rewards/accuracy_reward/mean": 0.09765625, "rewards/accuracy_reward/std": 0.29713961482048035, "step": 2518 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.787109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2043.0, "completions/mean_length": 1753.716796875, "completions/mean_terminated_length": 666.2201538085938, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8599470854314244, "frac_reward_zero_std": 0.8125, "grad_norm": 18.965083164635452, "learning_rate": 1.531275338325429e-07, "loss": 0.0109, "num_tokens": 2385863637.0, "reward": 0.05859375, "reward_std": 0.08413515239953995, "rewards/accuracy_reward/mean": 0.05859375, "rewards/accuracy_reward/std": 0.23509246110916138, "step": 2519 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.70703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1944.0, "completions/mean_length": 1642.27734375, "completions/mean_terminated_length": 663.1333618164062, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8602884697448152, "frac_reward_zero_std": 0.78125, "grad_norm": 24.364264106239652, "learning_rate": 1.528751135952585e-07, "loss": 0.0268, "num_tokens": 2386783107.0, "reward": 0.056640625, "reward_std": 0.08557207137346268, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 2520 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.705078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1801.0, "completions/mean_length": 1621.244140625, "completions/mean_terminated_length": 600.9867553710938, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.860629854058206, "frac_reward_zero_std": 0.65625, "grad_norm": 37.276246095158555, "learning_rate": 1.5262325700411534e-07, "loss": 0.0207, "num_tokens": 2387693648.0, "reward": 0.0859375, "reward_std": 0.1260918527841568, "rewards/accuracy_reward/mean": 0.0859375, "rewards/accuracy_reward/std": 0.28054583072662354, "step": 2521 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.759765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2015.0, "completions/mean_length": 1745.94921875, "completions/mean_terminated_length": 791.0650024414062, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8609712383715968, "frac_reward_zero_std": 0.71875, "grad_norm": 79.12150500995565, "learning_rate": 1.5237196441657767e-07, "loss": 0.0188, "num_tokens": 2388668902.0, "reward": 0.0390625, "reward_std": 0.09187985956668854, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 2522 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.705078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2030.0, "completions/mean_length": 1618.66015625, "completions/mean_terminated_length": 592.2251586914062, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8613126226849876, "frac_reward_zero_std": 0.71875, "grad_norm": 13.573922224854496, "learning_rate": 1.5212123618930924e-07, "loss": 0.0222, "num_tokens": 2389574200.0, "reward": 0.087890625, "reward_std": 0.11575096100568771, "rewards/accuracy_reward/mean": 0.087890625, "rewards/accuracy_reward/std": 0.2834126651287079, "step": 2523 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.724609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2025.0, "completions/mean_length": 1625.2890625, "completions/mean_terminated_length": 513.0496215820312, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8616540069983785, "frac_reward_zero_std": 0.59375, "grad_norm": 144.09319928591722, "learning_rate": 1.518710726781731e-07, "loss": 0.0222, "num_tokens": 2390486444.0, "reward": 0.115234375, "reward_std": 0.16641941666603088, "rewards/accuracy_reward/mean": 0.115234375, "rewards/accuracy_reward/std": 0.3196168541908264, "step": 2524 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.70703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2026.0, "completions/mean_length": 1646.890625, "completions/mean_terminated_length": 678.8800048828125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8619953913117693, "frac_reward_zero_std": 0.6875, "grad_norm": 75.02715437384968, "learning_rate": 1.5162147423823043e-07, "loss": 0.0129, "num_tokens": 2391402804.0, "reward": 0.08203125, "reward_std": 0.12615203857421875, "rewards/accuracy_reward/mean": 0.08203125, "rewards/accuracy_reward/std": 0.2746807038784027, "step": 2525 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.70703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1776.0, "completions/mean_length": 1623.779296875, "completions/mean_terminated_length": 600.3800048828125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.86233677562516, "frac_reward_zero_std": 0.78125, "grad_norm": 28.723624270462768, "learning_rate": 1.5137244122374076e-07, "loss": 0.0088, "num_tokens": 2392314467.0, "reward": 0.0703125, "reward_std": 0.0896916538476944, "rewards/accuracy_reward/mean": 0.0703125, "rewards/accuracy_reward/std": 0.25592297315597534, "step": 2526 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.751953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1779.0, "completions/mean_length": 1679.84375, "completions/mean_terminated_length": 563.779541015625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.8626781599385508, "frac_reward_zero_std": 0.90625, "grad_norm": 0.2915581539763728, "learning_rate": 1.5112397398816076e-07, "loss": 0.0092, "num_tokens": 2393255619.0, "reward": 0.013671875, "reward_std": 0.030584799125790596, "rewards/accuracy_reward/mean": 0.01411290280520916, "rewards/accuracy_reward/std": 0.11807557195425034, "step": 2527 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.818359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2041.0, "completions/mean_length": 1801.044921875, "completions/mean_terminated_length": 688.4193725585938, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.8630195442519416, "frac_reward_zero_std": 0.8125, "grad_norm": 64.31801265760994, "learning_rate": 1.508760728841442e-07, "loss": -0.0004, "num_tokens": 2394273466.0, "reward": 0.046875, "reward_std": 0.07494450360536575, "rewards/accuracy_reward/mean": 0.046875, "rewards/accuracy_reward/std": 0.21157780289649963, "step": 2528 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.728515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1793.0, "completions/mean_length": 1646.798828125, "completions/mean_terminated_length": 570.1942749023438, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8633609285653324, "frac_reward_zero_std": 0.8125, "grad_norm": 29.97471088726863, "learning_rate": 1.506287382635415e-07, "loss": 0.0087, "num_tokens": 2395192803.0, "reward": 0.052734375, "reward_std": 0.07285481691360474, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 2529 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.666015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2024.0, "completions/mean_length": 1570.09375, "completions/mean_terminated_length": 617.0760498046875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8637023128787232, "frac_reward_zero_std": 0.71875, "grad_norm": 23.7320107106578, "learning_rate": 1.503819704773987e-07, "loss": -0.0002, "num_tokens": 2396068515.0, "reward": 0.05859375, "reward_std": 0.10542967170476913, "rewards/accuracy_reward/mean": 0.060483869165182114, "rewards/accuracy_reward/std": 0.2386218160390854, "step": 2530 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2029.0, "completions/mean_length": 1630.630859375, "completions/mean_terminated_length": 678.173095703125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.864043697192114, "frac_reward_zero_std": 0.875, "grad_norm": 27.72477845038621, "learning_rate": 1.501357698759578e-07, "loss": 0.0051, "num_tokens": 2396978726.0, "reward": 0.060546875, "reward_std": 0.06095694378018379, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 2531 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.712890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1926.0, "completions/mean_length": 1640.2734375, "completions/mean_terminated_length": 627.89111328125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8643850815055049, "frac_reward_zero_std": 0.6875, "grad_norm": 26.18784945468839, "learning_rate": 1.498901368086553e-07, "loss": 0.0133, "num_tokens": 2397899538.0, "reward": 0.041015625, "reward_std": 0.09820909798145294, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 2532 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1705.0, "completions/mean_length": 1623.423828125, "completions/mean_terminated_length": 654.5192260742188, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.8647264658188957, "frac_reward_zero_std": 0.71875, "grad_norm": 47.10658308738134, "learning_rate": 1.4964507162412268e-07, "loss": -0.0061, "num_tokens": 2398802971.0, "reward": 0.05859375, "reward_std": 0.09919347614049911, "rewards/accuracy_reward/mean": 0.05859375, "rewards/accuracy_reward/std": 0.23509246110916138, "step": 2533 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.72265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.0, "completions/mean_length": 1659.291015625, "completions/mean_terminated_length": 646.457763671875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.8650678501322864, "frac_reward_zero_std": 0.65625, "grad_norm": 53.06636707655769, "learning_rate": 1.4940057467018482e-07, "loss": -0.0004, "num_tokens": 2399730896.0, "reward": 0.08984375, "reward_std": 0.13326609134674072, "rewards/accuracy_reward/mean": 0.08984375, "rewards/accuracy_reward/std": 0.2862374484539032, "step": 2534 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2036.0, "completions/mean_length": 1634.45703125, "completions/mean_terminated_length": 617.3648681640625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8654092344456772, "frac_reward_zero_std": 0.75, "grad_norm": 62.99446167026236, "learning_rate": 1.491566462938605e-07, "loss": 0.0155, "num_tokens": 2400654618.0, "reward": 0.03125, "reward_std": 0.08054865896701813, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 2535 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.740234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1875.0, "completions/mean_length": 1668.517578125, "completions/mean_terminated_length": 587.1353759765625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.865750618759068, "frac_reward_zero_std": 0.75, "grad_norm": 32.41018710459828, "learning_rate": 1.489132868413617e-07, "loss": 0.011, "num_tokens": 2401586403.0, "reward": 0.0703125, "reward_std": 0.10084541141986847, "rewards/accuracy_reward/mean": 0.0703125, "rewards/accuracy_reward/std": 0.25592297315597534, "step": 2536 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.708984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2027.0, "completions/mean_length": 1645.125, "completions/mean_terminated_length": 663.6241455078125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8660920030724588, "frac_reward_zero_std": 0.75, "grad_norm": 15.164600057614315, "learning_rate": 1.4867049665809232e-07, "loss": 0.0121, "num_tokens": 2402508771.0, "reward": 0.078125, "reward_std": 0.09011821448802948, "rewards/accuracy_reward/mean": 0.08064515888690948, "rewards/accuracy_reward/std": 0.2725643217563629, "step": 2537 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.77734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1929.0, "completions/mean_length": 1747.642578125, "completions/mean_terminated_length": 699.0263061523438, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.8664333873858496, "frac_reward_zero_std": 0.65625, "grad_norm": 51.61400381805593, "learning_rate": 1.4842827608864886e-07, "loss": 0.0307, "num_tokens": 2403482924.0, "reward": 0.125, "reward_std": 0.1436934620141983, "rewards/accuracy_reward/mean": 0.12903225421905518, "rewards/accuracy_reward/std": 0.33557409048080444, "step": 2538 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.681640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1994.0, "completions/mean_length": 1574.974609375, "completions/mean_terminated_length": 562.1779174804688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8667747716992404, "frac_reward_zero_std": 0.6875, "grad_norm": 64.91194608452355, "learning_rate": 1.48186625476819e-07, "loss": 0.014, "num_tokens": 2404367519.0, "reward": 0.060546875, "reward_std": 0.10074453055858612, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 2539 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.720703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1778.0, "completions/mean_length": 1650.431640625, "completions/mean_terminated_length": 624.5384521484375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8671161560126313, "frac_reward_zero_std": 0.8125, "grad_norm": 43.304218139688395, "learning_rate": 1.4794554516558166e-07, "loss": 0.0231, "num_tokens": 2405287612.0, "reward": 0.046875, "reward_std": 0.07340699434280396, "rewards/accuracy_reward/mean": 0.046875, "rewards/accuracy_reward/std": 0.21157780289649963, "step": 2540 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.689453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1967.0, "completions/mean_length": 1633.46484375, "completions/mean_terminated_length": 713.1446533203125, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.8674575403260221, "frac_reward_zero_std": 0.8125, "grad_norm": 26.235841971079545, "learning_rate": 1.477050354971061e-07, "loss": -0.0008, "num_tokens": 2406205754.0, "reward": 0.01171875, "reward_std": 0.046875, "rewards/accuracy_reward/mean": 0.01171875, "rewards/accuracy_reward/std": 0.10772226005792618, "step": 2541 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.720703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1923.0, "completions/mean_length": 1670.884765625, "completions/mean_terminated_length": 698.468505859375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8677989246394128, "frac_reward_zero_std": 0.8125, "grad_norm": 47.57754469776387, "learning_rate": 1.474650968127518e-07, "loss": -0.0035, "num_tokens": 2407149679.0, "reward": 0.015625, "reward_std": 0.051659777760505676, "rewards/accuracy_reward/mean": 0.015625, "rewards/accuracy_reward/std": 0.12414088100194931, "step": 2542 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.736328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1810.0, "completions/mean_length": 1665.06640625, "completions/mean_terminated_length": 595.6889038085938, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8681403089528036, "frac_reward_zero_std": 0.6875, "grad_norm": 115.63878532873551, "learning_rate": 1.4722572945306812e-07, "loss": 0.002, "num_tokens": 2408089649.0, "reward": 0.103515625, "reward_std": 0.1329134702682495, "rewards/accuracy_reward/mean": 0.103515625, "rewards/accuracy_reward/std": 0.30492907762527466, "step": 2543 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1776.0, "completions/mean_length": 1667.603515625, "completions/mean_terminated_length": 656.835693359375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8684816932661944, "frac_reward_zero_std": 0.90625, "grad_norm": 21.627427738617595, "learning_rate": 1.4698693375779296e-07, "loss": 0.0064, "num_tokens": 2409025430.0, "reward": 0.0078125, "reward_std": 0.02629890665411949, "rewards/accuracy_reward/mean": 0.0078125, "rewards/accuracy_reward/std": 0.08812850713729858, "step": 2544 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1970.0, "completions/mean_length": 1651.248046875, "completions/mean_terminated_length": 597.0214233398438, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8688230775795852, "frac_reward_zero_std": 0.8125, "grad_norm": 72.38246405724892, "learning_rate": 1.4674871006585326e-07, "loss": 0.0071, "num_tokens": 2409941989.0, "reward": 0.025390625, "reward_std": 0.05793476477265358, "rewards/accuracy_reward/mean": 0.025390625, "rewards/accuracy_reward/std": 0.15746226906776428, "step": 2545 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1975.0, "completions/mean_length": 1652.400390625, "completions/mean_terminated_length": 679.439208984375, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.869164461892976, "frac_reward_zero_std": 0.8125, "grad_norm": 29.746942194299564, "learning_rate": 1.46511058715364e-07, "loss": 0.0068, "num_tokens": 2410874594.0, "reward": 0.03125, "reward_std": 0.06558094918727875, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 2546 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2007.0, "completions/mean_length": 1657.62109375, "completions/mean_terminated_length": 578.3382568359375, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.8695058462063668, "frac_reward_zero_std": 0.84375, "grad_norm": 12.019235556000336, "learning_rate": 1.4627398004362774e-07, "loss": 0.0076, "num_tokens": 2411803520.0, "reward": 0.021484375, "reward_std": 0.05287160724401474, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 2547 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.716796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2031.0, "completions/mean_length": 1624.291015625, "completions/mean_terminated_length": 551.8689575195312, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.8698472305197577, "frac_reward_zero_std": 0.78125, "grad_norm": 21.472414184119458, "learning_rate": 1.4603747438713426e-07, "loss": -0.0065, "num_tokens": 2412713861.0, "reward": 0.064453125, "reward_std": 0.08725585043430328, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 2548 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.634765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1950.0, "completions/mean_length": 1524.498046875, "completions/mean_terminated_length": 614.66845703125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8701886148331485, "frac_reward_zero_std": 0.71875, "grad_norm": 61.347934229940904, "learning_rate": 1.458015420815601e-07, "loss": 0.0112, "num_tokens": 2413578116.0, "reward": 0.052734375, "reward_std": 0.09653191268444061, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 2549 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.705078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1996.0, "completions/mean_length": 1626.38671875, "completions/mean_terminated_length": 618.423828125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.8705299991465392, "frac_reward_zero_std": 0.75, "grad_norm": 33.136947453045245, "learning_rate": 1.4556618346176813e-07, "loss": -0.0029, "num_tokens": 2414490810.0, "reward": 0.06640625, "reward_std": 0.10300742089748383, "rewards/accuracy_reward/mean": 0.06640625, "rewards/accuracy_reward/std": 0.2492343932390213, "step": 2550 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2030.0, "completions/mean_length": 1584.25390625, "completions/mean_terminated_length": 564.0125122070312, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.87087138345993, "frac_reward_zero_std": 0.6875, "grad_norm": 16.613480975147798, "learning_rate": 1.453313988618067e-07, "loss": 0.0278, "num_tokens": 2415366460.0, "reward": 0.06640625, "reward_std": 0.10124340653419495, "rewards/accuracy_reward/mean": 0.06640625, "rewards/accuracy_reward/std": 0.2492343932390213, "step": 2551 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1932.0, "completions/mean_length": 1664.92578125, "completions/mean_terminated_length": 647.3928833007812, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.8712127677733208, "frac_reward_zero_std": 0.71875, "grad_norm": 42.12043570543552, "learning_rate": 1.4509718861490983e-07, "loss": 0.0288, "num_tokens": 2416302118.0, "reward": 0.072265625, "reward_std": 0.11970524489879608, "rewards/accuracy_reward/mean": 0.072265625, "rewards/accuracy_reward/std": 0.2591804563999176, "step": 2552 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.70703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2036.0, "completions/mean_length": 1654.19921875, "completions/mean_terminated_length": 703.82666015625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8715541520867116, "frac_reward_zero_std": 0.875, "grad_norm": 5.943678718648693, "learning_rate": 1.4486355305349583e-07, "loss": 0.0171, "num_tokens": 2417226860.0, "reward": 0.01953125, "reward_std": 0.04505910724401474, "rewards/accuracy_reward/mean": 0.01953125, "rewards/accuracy_reward/std": 0.1385180652141571, "step": 2553 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.70703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1917.0, "completions/mean_length": 1631.265625, "completions/mean_terminated_length": 625.5466918945312, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8718955364001024, "frac_reward_zero_std": 0.78125, "grad_norm": 40.84797510998624, "learning_rate": 1.4463049250916792e-07, "loss": 0.013, "num_tokens": 2418145748.0, "reward": 0.06640625, "reward_std": 0.09205513447523117, "rewards/accuracy_reward/mean": 0.06640625, "rewards/accuracy_reward/std": 0.2492343932390213, "step": 2554 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2018.0, "completions/mean_length": 1679.697265625, "completions/mean_terminated_length": 756.4177856445312, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.8722369207134932, "frac_reward_zero_std": 0.8125, "grad_norm": 78.3288705860024, "learning_rate": 1.4439800731271267e-07, "loss": 0.0047, "num_tokens": 2419088041.0, "reward": 0.033203125, "reward_std": 0.06755761057138443, "rewards/accuracy_reward/mean": 0.033203125, "rewards/accuracy_reward/std": 0.17934183776378632, "step": 2555 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.748046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1957.0, "completions/mean_length": 1668.708984375, "completions/mean_terminated_length": 542.596923828125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.872578305026884, "frac_reward_zero_std": 0.625, "grad_norm": 58.931756272459666, "learning_rate": 1.4416609779410049e-07, "loss": 0.0227, "num_tokens": 2420028740.0, "reward": 0.09375, "reward_std": 0.13984671235084534, "rewards/accuracy_reward/mean": 0.09677419066429138, "rewards/accuracy_reward/std": 0.2959485352039337, "step": 2556 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.69140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1993.0, "completions/mean_length": 1634.37109375, "completions/mean_terminated_length": 707.6329345703125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8729196893402749, "frac_reward_zero_std": 0.65625, "grad_norm": 118.9299061542486, "learning_rate": 1.4393476428248442e-07, "loss": 0.0216, "num_tokens": 2420939122.0, "reward": 0.05078125, "reward_std": 0.11877964437007904, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 2557 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1913.0, "completions/mean_length": 1638.830078125, "completions/mean_terminated_length": 613.4794311523438, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8732610736536656, "frac_reward_zero_std": 0.71875, "grad_norm": 19.442090108449733, "learning_rate": 1.4370400710620017e-07, "loss": 0.017, "num_tokens": 2421855403.0, "reward": 0.05078125, "reward_std": 0.10519799590110779, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 2558 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.650390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2039.0, "completions/mean_length": 1557.6640625, "completions/mean_terminated_length": 645.474853515625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8736024579670564, "frac_reward_zero_std": 0.78125, "grad_norm": 26.246613710760492, "learning_rate": 1.4347382659276529e-07, "loss": 0.007, "num_tokens": 2422730047.0, "reward": 0.0390625, "reward_std": 0.08109389245510101, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 2559 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.751953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1927.0, "completions/mean_length": 1690.6171875, "completions/mean_terminated_length": 607.2125854492188, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8739438422804472, "frac_reward_zero_std": 0.625, "grad_norm": 44.743599172277605, "learning_rate": 1.4324422306887873e-07, "loss": 0.0246, "num_tokens": 2423676843.0, "reward": 0.09765625, "reward_std": 0.14750081300735474, "rewards/accuracy_reward/mean": 0.09765625, "rewards/accuracy_reward/std": 0.29713961482048035, "step": 2560 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.763671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1842.0, "completions/mean_length": 1738.49609375, "completions/mean_terminated_length": 738.3635864257812, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.874285226593838, "frac_reward_zero_std": 0.75, "grad_norm": 32.515682819638286, "learning_rate": 1.430151968604211e-07, "loss": 0.0201, "num_tokens": 2424639177.0, "reward": 0.08203125, "reward_std": 0.09777135401964188, "rewards/accuracy_reward/mean": 0.08203125, "rewards/accuracy_reward/std": 0.2746807038784027, "step": 2561 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2020.0, "completions/mean_length": 1604.173828125, "completions/mean_terminated_length": 662.3963012695312, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.8746266109072288, "frac_reward_zero_std": 0.6875, "grad_norm": 21.455049688208057, "learning_rate": 1.4278674829245282e-07, "loss": 0.0244, "num_tokens": 2425539090.0, "reward": 0.087890625, "reward_std": 0.12038947641849518, "rewards/accuracy_reward/mean": 0.087890625, "rewards/accuracy_reward/std": 0.2834126651287079, "step": 2562 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1895.0, "completions/mean_length": 1620.07421875, "completions/mean_terminated_length": 606.5657958984375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.8749679952206196, "frac_reward_zero_std": 0.78125, "grad_norm": 103.71023040941192, "learning_rate": 1.425588776892151e-07, "loss": -0.0003, "num_tokens": 2426447272.0, "reward": 0.08984375, "reward_std": 0.08832141757011414, "rewards/accuracy_reward/mean": 0.0927419364452362, "rewards/accuracy_reward/std": 0.2903633117675781, "step": 2563 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1862.0, "completions/mean_length": 1665.513671875, "completions/mean_terminated_length": 707.0684814453125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8753093795340104, "frac_reward_zero_std": 0.84375, "grad_norm": 52.748714641869284, "learning_rate": 1.423315853741285e-07, "loss": 0.0112, "num_tokens": 2427390847.0, "reward": 0.02734375, "reward_std": 0.057157501578330994, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 2564 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.673828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1897.0, "completions/mean_length": 1594.484375, "completions/mean_terminated_length": 657.5808715820312, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8756507638474013, "frac_reward_zero_std": 0.71875, "grad_norm": 52.669373645234714, "learning_rate": 1.4210487166979283e-07, "loss": 0.0136, "num_tokens": 2428283591.0, "reward": 0.0625, "reward_std": 0.11487960815429688, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.2422981858253479, "step": 2565 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2005.0, "completions/mean_length": 1682.97265625, "completions/mean_terminated_length": 632.1363525390625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.875992148160792, "frac_reward_zero_std": 0.75, "grad_norm": 186.6926399034019, "learning_rate": 1.4187873689798684e-07, "loss": 0.018, "num_tokens": 2429225849.0, "reward": 0.02734375, "reward_std": 0.07779236882925034, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 2566 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.634765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1831.0, "completions/mean_length": 1525.42578125, "completions/mean_terminated_length": 617.2085571289062, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8763335324741828, "frac_reward_zero_std": 0.65625, "grad_norm": 38.87229133279316, "learning_rate": 1.416531813796674e-07, "loss": 0.0384, "num_tokens": 2430085747.0, "reward": 0.111328125, "reward_std": 0.14585818350315094, "rewards/accuracy_reward/mean": 0.111328125, "rewards/accuracy_reward/std": 0.31484565138816833, "step": 2567 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.689453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1890.0, "completions/mean_length": 1616.32421875, "completions/mean_terminated_length": 657.9496459960938, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.8766749167875736, "frac_reward_zero_std": 0.78125, "grad_norm": 28.3617982027029, "learning_rate": 1.4142820543496936e-07, "loss": 0.0141, "num_tokens": 2430985737.0, "reward": 0.037109375, "reward_std": 0.07768725603818893, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 2568 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.677734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1614.0, "completions/mean_length": 1588.41796875, "completions/mean_terminated_length": 621.9030151367188, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8770163011009644, "frac_reward_zero_std": 0.75, "grad_norm": 5.259464741690593, "learning_rate": 1.4120380938320487e-07, "loss": 0.0259, "num_tokens": 2431869999.0, "reward": 0.0546875, "reward_std": 0.09341736882925034, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 2569 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.712890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.0, "completions/mean_length": 1646.712890625, "completions/mean_terminated_length": 650.3197021484375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8773576854143552, "frac_reward_zero_std": 0.71875, "grad_norm": 56.90335062692698, "learning_rate": 1.4097999354286316e-07, "loss": 0.0256, "num_tokens": 2432796780.0, "reward": 0.048828125, "reward_std": 0.10068464279174805, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 2570 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.744140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1865.0, "completions/mean_length": 1668.396484375, "completions/mean_terminated_length": 564.3587646484375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.877699069727746, "frac_reward_zero_std": 0.8125, "grad_norm": 10.67960608019363, "learning_rate": 1.4075675823160982e-07, "loss": 0.0017, "num_tokens": 2433720279.0, "reward": 0.033203125, "reward_std": 0.06536141037940979, "rewards/accuracy_reward/mean": 0.033203125, "rewards/accuracy_reward/std": 0.17934183776378632, "step": 2571 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2038.0, "completions/mean_length": 1652.53125, "completions/mean_terminated_length": 642.263916015625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8780404540411368, "frac_reward_zero_std": 0.84375, "grad_norm": 10.622973922019595, "learning_rate": 1.4053410376628647e-07, "loss": 0.0011, "num_tokens": 2434650695.0, "reward": 0.037109375, "reward_std": 0.06024399772286415, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 2572 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1788.0, "completions/mean_length": 1610.2578125, "completions/mean_terminated_length": 647.2250366210938, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.8783818383545277, "frac_reward_zero_std": 0.8125, "grad_norm": 27.888683357959476, "learning_rate": 1.403120304629106e-07, "loss": -0.0058, "num_tokens": 2435552203.0, "reward": 0.0390625, "reward_std": 0.06936796009540558, "rewards/accuracy_reward/mean": 0.04032257944345474, "rewards/accuracy_reward/std": 0.19691328704357147, "step": 2573 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.720703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1951.0, "completions/mean_length": 1631.005859375, "completions/mean_terminated_length": 555.3916015625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.8787232226679184, "frac_reward_zero_std": 0.75, "grad_norm": 35.48613274007586, "learning_rate": 1.4009053863667448e-07, "loss": 0.0093, "num_tokens": 2436467262.0, "reward": 0.044921875, "reward_std": 0.0922069400548935, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 2574 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.681640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1882.0, "completions/mean_length": 1592.49609375, "completions/mean_terminated_length": 617.2147216796875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.8790646069813092, "frac_reward_zero_std": 0.78125, "grad_norm": 146.72170005130323, "learning_rate": 1.3986962860194528e-07, "loss": 0.0105, "num_tokens": 2437359852.0, "reward": 0.0546875, "reward_std": 0.08855310082435608, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 2575 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.689453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1768.0, "completions/mean_length": 1608.451171875, "completions/mean_terminated_length": 632.5974731445312, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.8794059912947, "frac_reward_zero_std": 0.8125, "grad_norm": 23.59732952144052, "learning_rate": 1.396493006722645e-07, "loss": 0.0246, "num_tokens": 2438257075.0, "reward": 0.0546875, "reward_std": 0.07218398153781891, "rewards/accuracy_reward/mean": 0.05645161122083664, "rewards/accuracy_reward/std": 0.23102475702762604, "step": 2576 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.701171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1993.0, "completions/mean_length": 1651.1171875, "completions/mean_terminated_length": 719.8692626953125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8797473756080908, "frac_reward_zero_std": 0.6875, "grad_norm": 261.621415972833, "learning_rate": 1.3942955516034715e-07, "loss": 0.0175, "num_tokens": 2439178463.0, "reward": 0.068359375, "reward_std": 0.10837717354297638, "rewards/accuracy_reward/mean": 0.068359375, "rewards/accuracy_reward/std": 0.25260838866233826, "step": 2577 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.705078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2013.0, "completions/mean_length": 1656.314453125, "completions/mean_terminated_length": 720.25830078125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8800887599214816, "frac_reward_zero_std": 0.84375, "grad_norm": 24.462477399923486, "learning_rate": 1.3921039237808198e-07, "loss": 0.0148, "num_tokens": 2440117024.0, "reward": 0.0234375, "reward_std": 0.05479498207569122, "rewards/accuracy_reward/mean": 0.0234375, "rewards/accuracy_reward/std": 0.15143637359142303, "step": 2578 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.705078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2041.0, "completions/mean_length": 1626.298828125, "completions/mean_terminated_length": 618.1258544921875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8804301442348724, "frac_reward_zero_std": 0.75, "grad_norm": 101.11125030362852, "learning_rate": 1.3899181263653026e-07, "loss": -0.0097, "num_tokens": 2441026633.0, "reward": 0.0859375, "reward_std": 0.09270582348108292, "rewards/accuracy_reward/mean": 0.0859375, "rewards/accuracy_reward/std": 0.28054583072662354, "step": 2579 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.697265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2030.0, "completions/mean_length": 1630.314453125, "completions/mean_terminated_length": 668.290283203125, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.8807715285482632, "frac_reward_zero_std": 0.71875, "grad_norm": 16.90407167905349, "learning_rate": 1.3877381624592616e-07, "loss": 0.018, "num_tokens": 2441932122.0, "reward": 0.064453125, "reward_std": 0.10746511071920395, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 2580 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.759765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2013.0, "completions/mean_length": 1697.818359375, "completions/mean_terminated_length": 590.8292236328125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.881112912861654, "frac_reward_zero_std": 0.8125, "grad_norm": 7.255693142079199, "learning_rate": 1.3855640351567553e-07, "loss": 0.0104, "num_tokens": 2442887453.0, "reward": 0.02734375, "reward_std": 0.06079617515206337, "rewards/accuracy_reward/mean": 0.02822580561041832, "rewards/accuracy_reward/std": 0.1657845675945282, "step": 2581 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.716796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2008.0, "completions/mean_length": 1639.5078125, "completions/mean_terminated_length": 605.5999755859375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8814542971750448, "frac_reward_zero_std": 0.96875, "grad_norm": 0.04592259591068291, "learning_rate": 1.3833957475435613e-07, "loss": 0.0048, "num_tokens": 2443809649.0, "reward": 0.001953125, "reward_std": 0.0078125, "rewards/accuracy_reward/mean": 0.001953125, "rewards/accuracy_reward/std": 0.04419417306780815, "step": 2582 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1975.0, "completions/mean_length": 1607.474609375, "completions/mean_terminated_length": 673.0304565429688, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8817956814884356, "frac_reward_zero_std": 0.625, "grad_norm": 168.8098879434718, "learning_rate": 1.3812333026971663e-07, "loss": 0.0056, "num_tokens": 2444715348.0, "reward": 0.05078125, "reward_std": 0.12114076316356659, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 2583 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.72265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2020.0, "completions/mean_length": 1674.541015625, "completions/mean_terminated_length": 701.4436645507812, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8821370658018264, "frac_reward_zero_std": 0.59375, "grad_norm": 12.044235324881859, "learning_rate": 1.3790767036867645e-07, "loss": 0.0013, "num_tokens": 2445649193.0, "reward": 0.0703125, "reward_std": 0.14334052801132202, "rewards/accuracy_reward/mean": 0.0703125, "rewards/accuracy_reward/std": 0.25592297315597534, "step": 2584 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.65234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1967.0, "completions/mean_length": 1562.564453125, "completions/mean_terminated_length": 651.6910400390625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8824784501152172, "frac_reward_zero_std": 0.78125, "grad_norm": 39.24333446966561, "learning_rate": 1.3769259535732561e-07, "loss": 0.0294, "num_tokens": 2446520474.0, "reward": 0.05078125, "reward_std": 0.07972268760204315, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 2585 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1872.0, "completions/mean_length": 1657.4765625, "completions/mean_terminated_length": 678.4931640625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.882819834428608, "frac_reward_zero_std": 0.5625, "grad_norm": 12.129675679783324, "learning_rate": 1.374781055409235e-07, "loss": 0.0386, "num_tokens": 2447442910.0, "reward": 0.083984375, "reward_std": 0.1506405919790268, "rewards/accuracy_reward/mean": 0.083984375, "rewards/accuracy_reward/std": 0.2776356339454651, "step": 2586 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.685546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2010.0, "completions/mean_length": 1587.4765625, "completions/mean_terminated_length": 583.478271484375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8831612187419988, "frac_reward_zero_std": 0.78125, "grad_norm": 94.50477724956566, "learning_rate": 1.372642012238993e-07, "loss": 0.018, "num_tokens": 2448332258.0, "reward": 0.06640625, "reward_std": 0.09734337776899338, "rewards/accuracy_reward/mean": 0.06854838877916336, "rewards/accuracy_reward/std": 0.25293970108032227, "step": 2587 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.724609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1781.0, "completions/mean_length": 1662.3203125, "completions/mean_terminated_length": 647.5177001953125, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.8835026030553896, "frac_reward_zero_std": 0.8125, "grad_norm": 0.9545814130024881, "learning_rate": 1.3705088270985103e-07, "loss": 0.0173, "num_tokens": 2449257478.0, "reward": 0.0234375, "reward_std": 0.05754890665411949, "rewards/accuracy_reward/mean": 0.0234375, "rewards/accuracy_reward/std": 0.15143637359142303, "step": 2588 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2011.0, "completions/mean_length": 1605.177734375, "completions/mean_terminated_length": 630.96875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.8838439873687804, "frac_reward_zero_std": 0.8125, "grad_norm": 28.325213225573407, "learning_rate": 1.3683815030154538e-07, "loss": -0.0029, "num_tokens": 2450162497.0, "reward": 0.06640625, "reward_std": 0.0727282464504242, "rewards/accuracy_reward/mean": 0.06854838877916336, "rewards/accuracy_reward/std": 0.25293973088264465, "step": 2589 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.732421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1904.0, "completions/mean_length": 1654.89453125, "completions/mean_terminated_length": 579.211669921875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8841853716821712, "frac_reward_zero_std": 0.8125, "grad_norm": 38.9051811748272, "learning_rate": 1.3662600430091707e-07, "loss": 0.0054, "num_tokens": 2451091179.0, "reward": 0.02734375, "reward_std": 0.06492365896701813, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 2590 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1994.0, "completions/mean_length": 1562.1484375, "completions/mean_terminated_length": 601.7442016601562, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.884526755995562, "frac_reward_zero_std": 0.71875, "grad_norm": 99.74967040805672, "learning_rate": 1.3641444500906846e-07, "loss": 0.0209, "num_tokens": 2451960471.0, "reward": 0.09765625, "reward_std": 0.10045164078474045, "rewards/accuracy_reward/mean": 0.09765625, "rewards/accuracy_reward/std": 0.29713961482048035, "step": 2591 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.685546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2011.0, "completions/mean_length": 1606.28515625, "completions/mean_terminated_length": 643.2919311523438, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8848681403089528, "frac_reward_zero_std": 0.59375, "grad_norm": 94.91501332618131, "learning_rate": 1.3620347272626933e-07, "loss": 0.0181, "num_tokens": 2452852169.0, "reward": 0.109375, "reward_std": 0.15148106217384338, "rewards/accuracy_reward/mean": 0.109375, "rewards/accuracy_reward/std": 0.31241437792778015, "step": 2592 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.708984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2010.0, "completions/mean_length": 1648.7265625, "completions/mean_terminated_length": 676.0, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8852095246223436, "frac_reward_zero_std": 0.8125, "grad_norm": 81.26350004864784, "learning_rate": 1.359930877519562e-07, "loss": 0.014, "num_tokens": 2453772141.0, "reward": 0.013671875, "reward_std": 0.04973640665411949, "rewards/accuracy_reward/mean": 0.01411290280520916, "rewards/accuracy_reward/std": 0.11807556450366974, "step": 2593 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.67578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1864.0, "completions/mean_length": 1562.9375, "completions/mean_terminated_length": 551.903564453125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8855509089357344, "frac_reward_zero_std": 0.6875, "grad_norm": 21.468465891854425, "learning_rate": 1.3578329038473222e-07, "loss": 0.0287, "num_tokens": 2454645933.0, "reward": 0.10546875, "reward_std": 0.13625310361385345, "rewards/accuracy_reward/mean": 0.10546875, "rewards/accuracy_reward/std": 0.3074568510055542, "step": 2594 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1910.0, "completions/mean_length": 1643.763671875, "completions/mean_terminated_length": 610.7152709960938, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8858922932491252, "frac_reward_zero_std": 0.78125, "grad_norm": 91.10358357724911, "learning_rate": 1.355740809223662e-07, "loss": 0.0264, "num_tokens": 2455556692.0, "reward": 0.048828125, "reward_std": 0.07933682948350906, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 2595 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.759765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2048.0, "completions/mean_length": 1694.693359375, "completions/mean_terminated_length": 577.3251953125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.886233677562516, "frac_reward_zero_std": 0.625, "grad_norm": 9.031722356817268, "learning_rate": 1.3536545966179274e-07, "loss": 0.0237, "num_tokens": 2456496999.0, "reward": 0.03515625, "reward_std": 0.10998040437698364, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 2596 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.732421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2017.0, "completions/mean_length": 1659.37109375, "completions/mean_terminated_length": 595.6058349609375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8865750618759068, "frac_reward_zero_std": 0.8125, "grad_norm": 97.03339588721056, "learning_rate": 1.3515742689911166e-07, "loss": -0.0032, "num_tokens": 2457423269.0, "reward": 0.048828125, "reward_std": 0.07163181155920029, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 2597 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.646484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1892.0, "completions/mean_length": 1577.64453125, "completions/mean_terminated_length": 717.4917602539062, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8869164461892975, "frac_reward_zero_std": 0.71875, "grad_norm": 9.032317677306038, "learning_rate": 1.3494998292958725e-07, "loss": -0.0011, "num_tokens": 2458305743.0, "reward": 0.044921875, "reward_std": 0.09528100490570068, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 2598 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1871.0, "completions/mean_length": 1603.654296875, "completions/mean_terminated_length": 510.8040771484375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8872578305026884, "frac_reward_zero_std": 0.8125, "grad_norm": 7.387011818490193, "learning_rate": 1.3474312804764853e-07, "loss": 0.0004, "num_tokens": 2459204798.0, "reward": 0.052734375, "reward_std": 0.06695909798145294, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 2599 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1935.0, "completions/mean_length": 1503.6328125, "completions/mean_terminated_length": 625.9795532226562, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8875992148160792, "frac_reward_zero_std": 0.71875, "grad_norm": 34.552076243846884, "learning_rate": 1.345368625468879e-07, "loss": 0.0108, "num_tokens": 2460048114.0, "reward": 0.05078125, "reward_std": 0.10519799590110779, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 2600 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.66015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2016.0, "completions/mean_length": 1560.275390625, "completions/mean_terminated_length": 612.8563232421875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.88794059912947, "frac_reward_zero_std": 0.78125, "grad_norm": 4.611143230975061, "learning_rate": 1.3433118672006173e-07, "loss": 0.0063, "num_tokens": 2460920415.0, "reward": 0.03515625, "reward_std": 0.07662828266620636, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 2601 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2017.0, "completions/mean_length": 1698.681640625, "completions/mean_terminated_length": 693.0681762695312, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8882819834428608, "frac_reward_zero_std": 0.75, "grad_norm": 12.358095907107318, "learning_rate": 1.3412610085908912e-07, "loss": 0.0285, "num_tokens": 2461865180.0, "reward": 0.068359375, "reward_std": 0.09506140649318695, "rewards/accuracy_reward/mean": 0.07056451588869095, "rewards/accuracy_reward/std": 0.25635457038879395, "step": 2602 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1967.0, "completions/mean_length": 1710.029296875, "completions/mean_terminated_length": 652.508056640625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8886233677562516, "frac_reward_zero_std": 0.78125, "grad_norm": 14.946390105870801, "learning_rate": 1.3392160525505191e-07, "loss": 0.0104, "num_tokens": 2462815339.0, "reward": 0.041015625, "reward_std": 0.07328042387962341, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 2603 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.66796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1972.0, "completions/mean_length": 1592.13671875, "completions/mean_terminated_length": 675.0470581054688, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.8889647520696424, "frac_reward_zero_std": 0.75, "grad_norm": 21.693148858229854, "learning_rate": 1.3371770019819433e-07, "loss": 0.0397, "num_tokens": 2463702641.0, "reward": 0.080078125, "reward_std": 0.10896982997655869, "rewards/accuracy_reward/mean": 0.080078125, "rewards/accuracy_reward/std": 0.271679550409317, "step": 2604 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.73046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1893.0, "completions/mean_length": 1661.62890625, "completions/mean_terminated_length": 614.5072631835938, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8893061363830332, "frac_reward_zero_std": 0.65625, "grad_norm": 62.488641186456604, "learning_rate": 1.3351438597792218e-07, "loss": 0.0064, "num_tokens": 2464628339.0, "reward": 0.06640625, "reward_std": 0.12857332825660706, "rewards/accuracy_reward/mean": 0.06640625, "rewards/accuracy_reward/std": 0.2492343932390213, "step": 2605 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.66796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2021.0, "completions/mean_length": 1570.572265625, "completions/mean_terminated_length": 610.1000366210938, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.889647520696424, "frac_reward_zero_std": 0.6875, "grad_norm": 31.26239939018135, "learning_rate": 1.3331166288280295e-07, "loss": 0.0061, "num_tokens": 2465504680.0, "reward": 0.103515625, "reward_std": 0.1269710510969162, "rewards/accuracy_reward/mean": 0.103515625, "rewards/accuracy_reward/std": 0.30492907762527466, "step": 2606 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1890.0, "completions/mean_length": 1606.53125, "completions/mean_terminated_length": 599.076904296875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8899889050098148, "frac_reward_zero_std": 0.75, "grad_norm": 36.76547732652226, "learning_rate": 1.3310953120056488e-07, "loss": 0.0218, "num_tokens": 2466401272.0, "reward": 0.064453125, "reward_std": 0.09385748207569122, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 2607 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.685546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1809.0, "completions/mean_length": 1610.966796875, "completions/mean_terminated_length": 658.1801147460938, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8903302893232056, "frac_reward_zero_std": 0.8125, "grad_norm": 18.193858885167952, "learning_rate": 1.3290799121809702e-07, "loss": 0.0116, "num_tokens": 2467303351.0, "reward": 0.044921875, "reward_std": 0.06943464279174805, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 2608 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.669921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1893.0, "completions/mean_length": 1551.76171875, "completions/mean_terminated_length": 544.6035766601562, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.8906716736365964, "frac_reward_zero_std": 0.65625, "grad_norm": 40.61312553932107, "learning_rate": 1.3270704322144832e-07, "loss": 0.026, "num_tokens": 2468171437.0, "reward": 0.0703125, "reward_std": 0.11123858392238617, "rewards/accuracy_reward/mean": 0.0703125, "rewards/accuracy_reward/std": 0.25592297315597534, "step": 2609 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1980.0, "completions/mean_length": 1623.931640625, "completions/mean_terminated_length": 690.9812622070312, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8910130579499872, "frac_reward_zero_std": 0.78125, "grad_norm": 150.40773777458514, "learning_rate": 1.3250668749582782e-07, "loss": 0.0056, "num_tokens": 2469076618.0, "reward": 0.05078125, "reward_std": 0.08477335423231125, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 2610 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.70703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2021.0, "completions/mean_length": 1599.83984375, "completions/mean_terminated_length": 518.280029296875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.891354442263378, "frac_reward_zero_std": 0.78125, "grad_norm": 19.83605761959913, "learning_rate": 1.3230692432560403e-07, "loss": 0.0118, "num_tokens": 2469978424.0, "reward": 0.060546875, "reward_std": 0.08847981691360474, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 2611 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2043.0, "completions/mean_length": 1518.25390625, "completions/mean_terminated_length": 605.2872314453125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8916958265767688, "frac_reward_zero_std": 0.71875, "grad_norm": 46.46193918051186, "learning_rate": 1.321077539943039e-07, "loss": 0.0334, "num_tokens": 2470828314.0, "reward": 0.056640625, "reward_std": 0.10715978592634201, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 2612 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.716796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2041.0, "completions/mean_length": 1644.53515625, "completions/mean_terminated_length": 623.3517456054688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8920372108901596, "frac_reward_zero_std": 0.90625, "grad_norm": 3.2328445596171917, "learning_rate": 1.319091767846136e-07, "loss": -0.0086, "num_tokens": 2471753868.0, "reward": 0.015625, "reward_std": 0.03125, "rewards/accuracy_reward/mean": 0.015625, "rewards/accuracy_reward/std": 0.12414088100194931, "step": 2613 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.701171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2008.0, "completions/mean_length": 1618.69140625, "completions/mean_terminated_length": 611.3594970703125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8923785952035505, "frac_reward_zero_std": 0.78125, "grad_norm": 15.513461665661305, "learning_rate": 1.3171119297837686e-07, "loss": 0.0045, "num_tokens": 2472658318.0, "reward": 0.033203125, "reward_std": 0.07564390450716019, "rewards/accuracy_reward/mean": 0.03427419438958168, "rewards/accuracy_reward/std": 0.18211629986763, "step": 2614 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.740234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1677.0, "completions/mean_length": 1670.748046875, "completions/mean_terminated_length": 595.7218017578125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8927199795169412, "frac_reward_zero_std": 0.8125, "grad_norm": 21.5804476201088, "learning_rate": 1.3151380285659565e-07, "loss": 0.0243, "num_tokens": 2473589981.0, "reward": 0.056640625, "reward_std": 0.07965600490570068, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 2615 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.794921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1894.0, "completions/mean_length": 1755.296875, "completions/mean_terminated_length": 620.7238159179688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.893061363830332, "frac_reward_zero_std": 0.8125, "grad_norm": 37.49575633053217, "learning_rate": 1.3131700669942907e-07, "loss": 0.0061, "num_tokens": 2474571797.0, "reward": 0.0390625, "reward_std": 0.06999341398477554, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 2616 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.732421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1893.0, "completions/mean_length": 1660.486328125, "completions/mean_terminated_length": 599.7737426757812, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.8934027481437228, "frac_reward_zero_std": 0.84375, "grad_norm": 4.9263298705819425, "learning_rate": 1.3112080478619333e-07, "loss": 0.0013, "num_tokens": 2475496798.0, "reward": 0.037109375, "reward_std": 0.06069765239953995, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 2617 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.712890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1995.0, "completions/mean_length": 1641.166015625, "completions/mean_terminated_length": 631.0, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.8937441324571136, "frac_reward_zero_std": 0.75, "grad_norm": 61.07437299990088, "learning_rate": 1.309251973953612e-07, "loss": 0.0291, "num_tokens": 2476416867.0, "reward": 0.04296875, "reward_std": 0.0878089889883995, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 2618 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1917.0, "completions/mean_length": 1643.3203125, "completions/mean_terminated_length": 648.0270385742188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8940855167705044, "frac_reward_zero_std": 0.875, "grad_norm": 4.163058106177084, "learning_rate": 1.3073018480456148e-07, "loss": 0.0023, "num_tokens": 2477339863.0, "reward": 0.015625, "reward_std": 0.040274329483509064, "rewards/accuracy_reward/mean": 0.015625, "rewards/accuracy_reward/std": 0.12414088100194931, "step": 2619 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.66796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1959.0, "completions/mean_length": 1594.013671875, "completions/mean_terminated_length": 681.0352783203125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8944269010838952, "frac_reward_zero_std": 0.625, "grad_norm": 28.762413729651612, "learning_rate": 1.3053576729057902e-07, "loss": 0.0356, "num_tokens": 2478239342.0, "reward": 0.083984375, "reward_std": 0.14133691787719727, "rewards/accuracy_reward/mean": 0.083984375, "rewards/accuracy_reward/std": 0.2776356339454651, "step": 2620 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.740234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1853.0, "completions/mean_length": 1676.376953125, "completions/mean_terminated_length": 617.3909912109375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.894768285397286, "frac_reward_zero_std": 0.8125, "grad_norm": 9.086692749327831, "learning_rate": 1.3034194512935377e-07, "loss": 0.0108, "num_tokens": 2479177359.0, "reward": 0.03125, "reward_std": 0.06783141195774078, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 2621 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.72265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1942.0, "completions/mean_length": 1638.021484375, "completions/mean_terminated_length": 569.767578125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8951096697106768, "frac_reward_zero_std": 0.78125, "grad_norm": 60.892173662204, "learning_rate": 1.3014871859598092e-07, "loss": -0.0016, "num_tokens": 2480100122.0, "reward": 0.064453125, "reward_std": 0.09315156936645508, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 2622 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.720703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1982.0, "completions/mean_length": 1709.080078125, "completions/mean_terminated_length": 834.5244750976562, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8954510540240675, "frac_reward_zero_std": 0.71875, "grad_norm": 46.55378966705435, "learning_rate": 1.299560879647101e-07, "loss": 0.0145, "num_tokens": 2481051299.0, "reward": 0.044921875, "reward_std": 0.09217274188995361, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 2623 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.68359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2028.0, "completions/mean_length": 1602.63671875, "completions/mean_terminated_length": 640.43212890625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8957924383374584, "frac_reward_zero_std": 0.84375, "grad_norm": 16.744258406877986, "learning_rate": 1.2976405350894536e-07, "loss": 0.0148, "num_tokens": 2481945993.0, "reward": 0.0390625, "reward_std": 0.061308603733778, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 2624 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.716796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1895.0, "completions/mean_length": 1640.171875, "completions/mean_terminated_length": 607.94482421875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8961338226508492, "frac_reward_zero_std": 0.75, "grad_norm": 9.01510158767108, "learning_rate": 1.295726155012445e-07, "loss": 0.0091, "num_tokens": 2482862721.0, "reward": 0.048828125, "reward_std": 0.0921536535024643, "rewards/accuracy_reward/mean": 0.05040322616696358, "rewards/accuracy_reward/std": 0.21899642050266266, "step": 2625 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.716796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2037.0, "completions/mean_length": 1642.0234375, "completions/mean_terminated_length": 614.4827270507812, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.89647520696424, "frac_reward_zero_std": 0.6875, "grad_norm": 35.766057823413796, "learning_rate": 1.2938177421331875e-07, "loss": 0.0134, "num_tokens": 2483787229.0, "reward": 0.08984375, "reward_std": 0.12364465743303299, "rewards/accuracy_reward/mean": 0.08984375, "rewards/accuracy_reward/std": 0.2862374484539032, "step": 2626 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.658203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.0, "completions/mean_length": 1570.669921875, "completions/mean_terminated_length": 651.4685668945312, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.8968165912776308, "frac_reward_zero_std": 0.6875, "grad_norm": 170.32923708655812, "learning_rate": 1.2919152991603235e-07, "loss": 0.0117, "num_tokens": 2484666276.0, "reward": 0.05859375, "reward_std": 0.11527761071920395, "rewards/accuracy_reward/mean": 0.05859375, "rewards/accuracy_reward/std": 0.23509246110916138, "step": 2627 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.736328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2009.0, "completions/mean_length": 1685.908203125, "completions/mean_terminated_length": 674.7333374023438, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8971579755910216, "frac_reward_zero_std": 0.71875, "grad_norm": 19.84809704462691, "learning_rate": 1.2900188287940223e-07, "loss": 0.0058, "num_tokens": 2485603029.0, "reward": 0.04296875, "reward_std": 0.09338457882404327, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 2628 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.73046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2012.0, "completions/mean_length": 1677.45703125, "completions/mean_terminated_length": 673.2318725585938, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8974993599044124, "frac_reward_zero_std": 0.71875, "grad_norm": 12.473307080978197, "learning_rate": 1.2881283337259784e-07, "loss": 0.0062, "num_tokens": 2486532959.0, "reward": 0.05078125, "reward_std": 0.10292718559503555, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 2629 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.791015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1616.0, "completions/mean_length": 1735.9375, "completions/mean_terminated_length": 554.766357421875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8978407442178032, "frac_reward_zero_std": 0.84375, "grad_norm": 40.78364422485982, "learning_rate": 1.2862438166394022e-07, "loss": 0.0072, "num_tokens": 2487505519.0, "reward": 0.05078125, "reward_std": 0.0595787949860096, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 2630 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.69140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1982.0, "completions/mean_length": 1589.44140625, "completions/mean_terminated_length": 562.0379638671875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.898182128531194, "frac_reward_zero_std": 0.6875, "grad_norm": 25.5979100696351, "learning_rate": 1.284365280209022e-07, "loss": 0.0107, "num_tokens": 2488394705.0, "reward": 0.0625, "reward_std": 0.10975532233715057, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.2422981858253479, "step": 2631 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1944.0, "completions/mean_length": 1665.7421875, "completions/mean_terminated_length": 608.9117431640625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8985235128445848, "frac_reward_zero_std": 0.90625, "grad_norm": 25.11178383328119, "learning_rate": 1.2824927271010777e-07, "loss": -0.0016, "num_tokens": 2489326829.0, "reward": 0.03515625, "reward_std": 0.04318207502365112, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 2632 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.73828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1975.0, "completions/mean_length": 1662.48828125, "completions/mean_terminated_length": 575.0, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8988648971579756, "frac_reward_zero_std": 0.71875, "grad_norm": 154.26176219196702, "learning_rate": 1.2806261599733127e-07, "loss": 0.0301, "num_tokens": 2490256343.0, "reward": 0.068359375, "reward_std": 0.12042367458343506, "rewards/accuracy_reward/mean": 0.068359375, "rewards/accuracy_reward/std": 0.25260838866233826, "step": 2633 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1770.0, "completions/mean_length": 1630.689453125, "completions/mean_terminated_length": 584.5548095703125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.8992062814713664, "frac_reward_zero_std": 0.71875, "grad_norm": 50.20600807016281, "learning_rate": 1.278765581474981e-07, "loss": -0.0133, "num_tokens": 2491166456.0, "reward": 0.05859375, "reward_std": 0.09814241528511047, "rewards/accuracy_reward/mean": 0.05859375, "rewards/accuracy_reward/std": 0.23509246110916138, "step": 2634 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.673828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1617.0, "completions/mean_length": 1561.33984375, "completions/mean_terminated_length": 555.964111328125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.8995476657847572, "frac_reward_zero_std": 0.84375, "grad_norm": 8.533965691232433, "learning_rate": 1.276910994246831e-07, "loss": 0.0053, "num_tokens": 2492039974.0, "reward": 0.0546875, "reward_std": 0.052172206342220306, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 2635 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.716796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1924.0, "completions/mean_length": 1632.93359375, "completions/mean_terminated_length": 582.38623046875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.899889050098148, "frac_reward_zero_std": 0.9375, "grad_norm": 0.0892123151270677, "learning_rate": 1.275062400921112e-07, "loss": 0.0055, "num_tokens": 2492958772.0, "reward": 0.005859375, "reward_std": 0.01848640665411949, "rewards/accuracy_reward/mean": 0.006048386916518211, "rewards/accuracy_reward/std": 0.07761410623788834, "step": 2636 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.763671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2015.0, "completions/mean_length": 1717.1796875, "completions/mean_terminated_length": 648.165283203125, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.9002304344115388, "frac_reward_zero_std": 0.84375, "grad_norm": 17.465575480785812, "learning_rate": 1.273219804121562e-07, "loss": 0.0071, "num_tokens": 2493919536.0, "reward": 0.07421875, "reward_std": 0.0721839889883995, "rewards/accuracy_reward/mean": 0.07421875, "rewards/accuracy_reward/std": 0.2623828947544098, "step": 2637 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.68359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1914.0, "completions/mean_length": 1612.67578125, "completions/mean_terminated_length": 672.1605224609375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9005718187249296, "frac_reward_zero_std": 0.65625, "grad_norm": 76.31475256587117, "learning_rate": 1.2713832064634125e-07, "loss": 0.0051, "num_tokens": 2494823994.0, "reward": 0.0546875, "reward_std": 0.11803390085697174, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 2638 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.68359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2019.0, "completions/mean_length": 1602.498046875, "completions/mean_terminated_length": 639.9938354492188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9009132030383203, "frac_reward_zero_std": 0.75, "grad_norm": 13.583995743047526, "learning_rate": 1.2695526105533768e-07, "loss": 0.0049, "num_tokens": 2495726009.0, "reward": 0.048828125, "reward_std": 0.09550845623016357, "rewards/accuracy_reward/mean": 0.05040322616696358, "rewards/accuracy_reward/std": 0.21899642050266266, "step": 2639 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.748046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1993.0, "completions/mean_length": 1697.796875, "completions/mean_terminated_length": 658.8062133789062, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9012545873517112, "frac_reward_zero_std": 0.71875, "grad_norm": 41.06824188920856, "learning_rate": 1.26772801898965e-07, "loss": 0.0185, "num_tokens": 2496677569.0, "reward": 0.0703125, "reward_std": 0.11607901006937027, "rewards/accuracy_reward/mean": 0.07258064299821854, "rewards/accuracy_reward/std": 0.25970885157585144, "step": 2640 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.697265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2036.0, "completions/mean_length": 1605.6640625, "completions/mean_terminated_length": 586.864501953125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.901595971665102, "frac_reward_zero_std": 0.75, "grad_norm": 46.015264450247926, "learning_rate": 1.2659094343619087e-07, "loss": 0.0154, "num_tokens": 2497574901.0, "reward": 0.056640625, "reward_std": 0.08395528793334961, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 2641 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.759765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1982.0, "completions/mean_length": 1724.259765625, "completions/mean_terminated_length": 700.3983764648438, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9019373559784928, "frac_reward_zero_std": 0.875, "grad_norm": 14.438538727672466, "learning_rate": 1.2640968592512978e-07, "loss": -0.006, "num_tokens": 2498536506.0, "reward": 0.029296875, "reward_std": 0.04478531330823898, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 2642 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.70703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2019.0, "completions/mean_length": 1615.744140625, "completions/mean_terminated_length": 572.8400268554688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9022787402918836, "frac_reward_zero_std": 0.5625, "grad_norm": 168.58015065708898, "learning_rate": 1.2622902962304394e-07, "loss": 0.0058, "num_tokens": 2499447991.0, "reward": 0.072265625, "reward_std": 0.15269054472446442, "rewards/accuracy_reward/mean": 0.07459677755832672, "rewards/accuracy_reward/std": 0.263004869222641, "step": 2643 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.685546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1936.0, "completions/mean_length": 1588.134765625, "completions/mean_terminated_length": 585.5714111328125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9026201246052744, "frac_reward_zero_std": 0.78125, "grad_norm": 13.140874581705576, "learning_rate": 1.2604897478634172e-07, "loss": 0.0096, "num_tokens": 2500342092.0, "reward": 0.03515625, "reward_std": 0.06085042655467987, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 2644 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1684.0, "completions/mean_length": 1587.537109375, "completions/mean_terminated_length": 610.457275390625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9029615089186652, "frac_reward_zero_std": 0.78125, "grad_norm": 39.653449839858204, "learning_rate": 1.2586952167057805e-07, "loss": 0.0181, "num_tokens": 2501229039.0, "reward": 0.056640625, "reward_std": 0.0815330371260643, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 2645 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.66015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1912.0, "completions/mean_length": 1551.962890625, "completions/mean_terminated_length": 588.7069091796875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.903302893232056, "frac_reward_zero_std": 0.84375, "grad_norm": 27.250919203736725, "learning_rate": 1.256906705304539e-07, "loss": 0.01, "num_tokens": 2502107020.0, "reward": 0.03125, "reward_std": 0.05969182401895523, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 2646 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.708984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1721.0, "completions/mean_length": 1619.638671875, "completions/mean_terminated_length": 576.0469970703125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9036442775454467, "frac_reward_zero_std": 0.65625, "grad_norm": 158.77926796001358, "learning_rate": 1.2551242161981563e-07, "loss": 0.0253, "num_tokens": 2503012787.0, "reward": 0.083984375, "reward_std": 0.1171422079205513, "rewards/accuracy_reward/mean": 0.08669354766607285, "rewards/accuracy_reward/std": 0.281669557094574, "step": 2647 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1826.0, "completions/mean_length": 1575.03515625, "completions/mean_terminated_length": 640.4302368164062, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.9039856618588376, "frac_reward_zero_std": 0.78125, "grad_norm": 36.63063415437126, "learning_rate": 1.253347751916551e-07, "loss": 0.0249, "num_tokens": 2503901381.0, "reward": 0.044921875, "reward_std": 0.08298446238040924, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 2648 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.693359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1865.0, "completions/mean_length": 1585.76953125, "completions/mean_terminated_length": 540.5987548828125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.9043270461722284, "frac_reward_zero_std": 0.78125, "grad_norm": 15.716262182459976, "learning_rate": 1.2515773149810875e-07, "loss": 0.0282, "num_tokens": 2504795439.0, "reward": 0.048828125, "reward_std": 0.07988205552101135, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 2649 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1952.0, "completions/mean_length": 1658.484375, "completions/mean_terminated_length": 663.0555419921875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9046684304856192, "frac_reward_zero_std": 0.78125, "grad_norm": 47.247834642059466, "learning_rate": 1.24981290790458e-07, "loss": 0.0118, "num_tokens": 2505717591.0, "reward": 0.05078125, "reward_std": 0.08501990139484406, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 2650 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1944.0, "completions/mean_length": 1604.31640625, "completions/mean_terminated_length": 591.8076782226562, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.90500981479901, "frac_reward_zero_std": 0.5625, "grad_norm": 39.76025085675048, "learning_rate": 1.2480545331912786e-07, "loss": -0.0214, "num_tokens": 2506610393.0, "reward": 0.119140625, "reward_std": 0.159950852394104, "rewards/accuracy_reward/mean": 0.119140625, "rewards/accuracy_reward/std": 0.32427072525024414, "step": 2651 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.619140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2020.0, "completions/mean_length": 1476.30078125, "completions/mean_terminated_length": 546.923095703125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.9053511991124008, "frac_reward_zero_std": 0.875, "grad_norm": 17.72578038485607, "learning_rate": 1.246302193336876e-07, "loss": 0.0079, "num_tokens": 2507443827.0, "reward": 0.025390625, "reward_std": 0.04803257808089256, "rewards/accuracy_reward/mean": 0.025390625, "rewards/accuracy_reward/std": 0.15746226906776428, "step": 2652 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.681640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1997.0, "completions/mean_length": 1586.63671875, "completions/mean_terminated_length": 598.809814453125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.9056925834257916, "frac_reward_zero_std": 0.75, "grad_norm": 20.682672032273622, "learning_rate": 1.2445558908284983e-07, "loss": 0.0115, "num_tokens": 2508332889.0, "reward": 0.03515625, "reward_std": 0.0853334367275238, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 2653 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.63671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2011.0, "completions/mean_length": 1523.498046875, "completions/mean_terminated_length": 604.2096557617188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9060339677391824, "frac_reward_zero_std": 0.65625, "grad_norm": 45.37614648473046, "learning_rate": 1.2428156281447017e-07, "loss": 0.0218, "num_tokens": 2509180072.0, "reward": 0.06640625, "reward_std": 0.13144925236701965, "rewards/accuracy_reward/mean": 0.06640625, "rewards/accuracy_reward/std": 0.2492343932390213, "step": 2654 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2032.0, "completions/mean_length": 1656.8203125, "completions/mean_terminated_length": 730.625, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.9063753520525731, "frac_reward_zero_std": 0.78125, "grad_norm": 13.082394539660557, "learning_rate": 1.2410814077554717e-07, "loss": 0.0085, "num_tokens": 2510108940.0, "reward": 0.06640625, "reward_std": 0.0961792916059494, "rewards/accuracy_reward/mean": 0.06640625, "rewards/accuracy_reward/std": 0.2492343932390213, "step": 2655 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.75390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1930.0, "completions/mean_length": 1688.666015625, "completions/mean_terminated_length": 587.8492431640625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.906716736365964, "frac_reward_zero_std": 0.71875, "grad_norm": 72.0515138179927, "learning_rate": 1.239353232122216e-07, "loss": 0.02, "num_tokens": 2511041393.0, "reward": 0.076171875, "reward_std": 0.0986412987112999, "rewards/accuracy_reward/mean": 0.076171875, "rewards/accuracy_reward/std": 0.26553234457969666, "step": 2656 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.705078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2034.0, "completions/mean_length": 1648.125, "completions/mean_terminated_length": 692.52978515625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9070581206793548, "frac_reward_zero_std": 0.75, "grad_norm": 40.247496353478766, "learning_rate": 1.2376311036977652e-07, "loss": -0.0018, "num_tokens": 2511964465.0, "reward": 0.0546875, "reward_std": 0.10276086628437042, "rewards/accuracy_reward/mean": 0.05645161122083664, "rewards/accuracy_reward/std": 0.23102475702762604, "step": 2657 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.697265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2007.0, "completions/mean_length": 1605.15625, "completions/mean_terminated_length": 585.1870727539062, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9073995049927456, "frac_reward_zero_std": 0.6875, "grad_norm": 46.041308995122705, "learning_rate": 1.2359150249263649e-07, "loss": 0.0153, "num_tokens": 2512862785.0, "reward": 0.080078125, "reward_std": 0.11981005221605301, "rewards/accuracy_reward/mean": 0.080078125, "rewards/accuracy_reward/std": 0.271679550409317, "step": 2658 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1628.0, "completions/mean_length": 1670.048828125, "completions/mean_terminated_length": 625.125, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.9077408893061364, "frac_reward_zero_std": 0.6875, "grad_norm": 15.350329247527585, "learning_rate": 1.2342049982436734e-07, "loss": 0.0234, "num_tokens": 2513794330.0, "reward": 0.0546875, "reward_std": 0.11300258338451385, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 2659 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.759765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1743.0, "completions/mean_length": 1687.546875, "completions/mean_terminated_length": 547.5772094726562, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9080822736195272, "frac_reward_zero_std": 0.84375, "grad_norm": 10.41944089934865, "learning_rate": 1.2325010260767639e-07, "loss": -0.002, "num_tokens": 2514729954.0, "reward": 0.017578125, "reward_std": 0.048086829483509064, "rewards/accuracy_reward/mean": 0.017578125, "rewards/accuracy_reward/std": 0.13154059648513794, "step": 2660 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.697265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1826.0, "completions/mean_length": 1600.4609375, "completions/mean_terminated_length": 569.6774291992188, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.908423657932918, "frac_reward_zero_std": 0.84375, "grad_norm": 11.967166108161393, "learning_rate": 1.2308031108441105e-07, "loss": 0.0065, "num_tokens": 2515621982.0, "reward": 0.015625, "reward_std": 0.04522542655467987, "rewards/accuracy_reward/mean": 0.015625, "rewards/accuracy_reward/std": 0.12414088100194931, "step": 2661 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.693359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2032.0, "completions/mean_length": 1668.9921875, "completions/mean_terminated_length": 812.3312377929688, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9087650422463088, "frac_reward_zero_std": 0.65625, "grad_norm": 83.27295747751994, "learning_rate": 1.2291112549555952e-07, "loss": 0.0142, "num_tokens": 2516555258.0, "reward": 0.08203125, "reward_std": 0.12404991686344147, "rewards/accuracy_reward/mean": 0.08203125, "rewards/accuracy_reward/std": 0.2746807038784027, "step": 2662 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.68359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1971.0, "completions/mean_length": 1610.734375, "completions/mean_terminated_length": 666.0247192382812, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.9091064265596995, "frac_reward_zero_std": 0.75, "grad_norm": 41.31636336103393, "learning_rate": 1.2274254608124973e-07, "loss": -0.0033, "num_tokens": 2517459170.0, "reward": 0.0390625, "reward_std": 0.08510598540306091, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 2663 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.654296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1850.0, "completions/mean_length": 1534.189453125, "completions/mean_terminated_length": 561.72314453125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9094478108730903, "frac_reward_zero_std": 0.75, "grad_norm": 39.37147980159843, "learning_rate": 1.2257457308074925e-07, "loss": 0.0095, "num_tokens": 2518316739.0, "reward": 0.095703125, "reward_std": 0.09121561795473099, "rewards/accuracy_reward/mean": 0.095703125, "rewards/accuracy_reward/std": 0.2944713830947876, "step": 2664 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.69140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1900.0, "completions/mean_length": 1595.177734375, "completions/mean_terminated_length": 580.6265869140625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9097891951864812, "frac_reward_zero_std": 0.8125, "grad_norm": 39.55892785541511, "learning_rate": 1.2240720673246515e-07, "loss": 0.0129, "num_tokens": 2519217310.0, "reward": 0.052734375, "reward_std": 0.06750431656837463, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 2665 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.74609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1978.0, "completions/mean_length": 1698.146484375, "completions/mean_terminated_length": 670.1153564453125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.910130579499872, "frac_reward_zero_std": 0.8125, "grad_norm": 20.896062231083587, "learning_rate": 1.2224044727394326e-07, "loss": 0.0281, "num_tokens": 2520161273.0, "reward": 0.02734375, "reward_std": 0.06546888500452042, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 2666 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.69140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1985.0, "completions/mean_length": 1611.3515625, "completions/mean_terminated_length": 633.0379638671875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9104719638132628, "frac_reward_zero_std": 0.6875, "grad_norm": 179.0401877357721, "learning_rate": 1.2207429494186826e-07, "loss": 0.0239, "num_tokens": 2521058765.0, "reward": 0.06640625, "reward_std": 0.11439286172389984, "rewards/accuracy_reward/mean": 0.06640625, "rewards/accuracy_reward/std": 0.2492343932390213, "step": 2667 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.693359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1923.0, "completions/mean_length": 1615.880859375, "completions/mean_terminated_length": 638.7962036132812, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9108133481266536, "frac_reward_zero_std": 0.84375, "grad_norm": 18.719464381534646, "learning_rate": 1.219087499720628e-07, "loss": 0.012, "num_tokens": 2521968784.0, "reward": 0.046875, "reward_std": 0.06942769140005112, "rewards/accuracy_reward/mean": 0.04838709533214569, "rewards/accuracy_reward/std": 0.21479946374893188, "step": 2668 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.732421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1803.0, "completions/mean_length": 1644.283203125, "completions/mean_terminated_length": 539.218994140625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.9111547324400444, "frac_reward_zero_std": 0.8125, "grad_norm": 20.655388367207642, "learning_rate": 1.2174381259948785e-07, "loss": -0.0066, "num_tokens": 2522900417.0, "reward": 0.037109375, "reward_std": 0.06767063587903976, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 2669 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2019.0, "completions/mean_length": 1668.72265625, "completions/mean_terminated_length": 577.2651977539062, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.9114961167534352, "frac_reward_zero_std": 0.71875, "grad_norm": 37.62800210261315, "learning_rate": 1.2157948305824184e-07, "loss": 0.0263, "num_tokens": 2523833507.0, "reward": 0.052734375, "reward_std": 0.09448197484016418, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 2670 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1897.0, "completions/mean_length": 1583.7109375, "completions/mean_terminated_length": 598.5121459960938, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.9118375010668259, "frac_reward_zero_std": 0.75, "grad_norm": 9.313348720614254, "learning_rate": 1.2141576158156031e-07, "loss": 0.0293, "num_tokens": 2524719647.0, "reward": 0.0546875, "reward_std": 0.09193411469459534, "rewards/accuracy_reward/mean": 0.05645161122083664, "rewards/accuracy_reward/std": 0.23102475702762604, "step": 2671 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.72265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2040.0, "completions/mean_length": 1660.62109375, "completions/mean_terminated_length": 651.2534790039062, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.9121788853802167, "frac_reward_zero_std": 0.78125, "grad_norm": 37.55590370457818, "learning_rate": 1.2125264840181623e-07, "loss": 0.0187, "num_tokens": 2525643981.0, "reward": 0.037109375, "reward_std": 0.08010855317115784, "rewards/accuracy_reward/mean": 0.038306452333927155, "rewards/accuracy_reward/std": 0.19212885200977325, "step": 2672 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1833.0, "completions/mean_length": 1687.822265625, "completions/mean_terminated_length": 650.9469604492188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9125202696936076, "frac_reward_zero_std": 0.84375, "grad_norm": 8.962958414511192, "learning_rate": 1.2109014375051868e-07, "loss": -0.004, "num_tokens": 2526582706.0, "reward": 0.046875, "reward_std": 0.056411758065223694, "rewards/accuracy_reward/mean": 0.046875, "rewards/accuracy_reward/std": 0.21157780289649963, "step": 2673 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2039.0, "completions/mean_length": 1573.078125, "completions/mean_terminated_length": 600.8988037109375, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.9128616540069984, "frac_reward_zero_std": 0.78125, "grad_norm": 20.049250347128748, "learning_rate": 1.2092824785831342e-07, "loss": 0.0221, "num_tokens": 2527470810.0, "reward": 0.060546875, "reward_std": 0.0706586092710495, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 2674 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.69140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1999.0, "completions/mean_length": 1618.92578125, "completions/mean_terminated_length": 657.582275390625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9132030383203892, "frac_reward_zero_std": 0.59375, "grad_norm": 30.752971624082775, "learning_rate": 1.2076696095498203e-07, "loss": 0.0067, "num_tokens": 2528381556.0, "reward": 0.10546875, "reward_std": 0.1424015313386917, "rewards/accuracy_reward/mean": 0.10546875, "rewards/accuracy_reward/std": 0.3074568510055542, "step": 2675 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2045.0, "completions/mean_length": 1591.306640625, "completions/mean_terminated_length": 622.2255859375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.91354442263378, "frac_reward_zero_std": 0.84375, "grad_norm": 14.31009969927499, "learning_rate": 1.2060628326944175e-07, "loss": 0.0103, "num_tokens": 2529274353.0, "reward": 0.04296875, "reward_std": 0.059520021080970764, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 2676 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2040.0, "completions/mean_length": 1654.22265625, "completions/mean_terminated_length": 648.2569580078125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9138858069471708, "frac_reward_zero_std": 0.625, "grad_norm": 29.36820642916589, "learning_rate": 1.204462150297452e-07, "loss": 0.0091, "num_tokens": 2530207923.0, "reward": 0.08984375, "reward_std": 0.13106893002986908, "rewards/accuracy_reward/mean": 0.08984375, "rewards/accuracy_reward/std": 0.2862374484539032, "step": 2677 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.76171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1964.0, "completions/mean_length": 1711.08203125, "completions/mean_terminated_length": 634.0491333007812, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.9142271912605616, "frac_reward_zero_std": 0.75, "grad_norm": 33.145445098029626, "learning_rate": 1.202867564630799e-07, "loss": 0.0275, "num_tokens": 2531165709.0, "reward": 0.064453125, "reward_std": 0.09414386004209518, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 2678 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.68359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1805.0, "completions/mean_length": 1592.2890625, "completions/mean_terminated_length": 607.7283935546875, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.9145685755739523, "frac_reward_zero_std": 0.65625, "grad_norm": 65.36714056252808, "learning_rate": 1.2012790779576833e-07, "loss": 0.0214, "num_tokens": 2532060193.0, "reward": 0.064453125, "reward_std": 0.12240587919950485, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 2679 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.74609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2036.0, "completions/mean_length": 1691.6015625, "completions/mean_terminated_length": 644.3384399414062, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.9149099598873431, "frac_reward_zero_std": 0.6875, "grad_norm": 62.036618332383505, "learning_rate": 1.1996966925326677e-07, "loss": 0.0208, "num_tokens": 2533008581.0, "reward": 0.056640625, "reward_std": 0.1190534457564354, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 2680 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.712890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1897.0, "completions/mean_length": 1625.595703125, "completions/mean_terminated_length": 577.1564331054688, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.915251344200734, "frac_reward_zero_std": 0.78125, "grad_norm": 18.0400499286434, "learning_rate": 1.1981204106016626e-07, "loss": 0.0296, "num_tokens": 2533930246.0, "reward": 0.052734375, "reward_std": 0.08285094797611237, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 2681 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1947.0, "completions/mean_length": 1637.884765625, "completions/mean_terminated_length": 629.2229614257812, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.9155927285141248, "frac_reward_zero_std": 0.8125, "grad_norm": 50.519699906866606, "learning_rate": 1.196550234401909e-07, "loss": -0.0009, "num_tokens": 2534849035.0, "reward": 0.017578125, "reward_std": 0.05452118441462517, "rewards/accuracy_reward/mean": 0.017578125, "rewards/accuracy_reward/std": 0.13154059648513794, "step": 2682 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1840.0, "completions/mean_length": 1542.904296875, "completions/mean_terminated_length": 642.5162963867188, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9159341128275156, "frac_reward_zero_std": 0.78125, "grad_norm": 26.154366962389492, "learning_rate": 1.194986166161986e-07, "loss": 0.0184, "num_tokens": 2535710554.0, "reward": 0.05859375, "reward_std": 0.0921671986579895, "rewards/accuracy_reward/mean": 0.05859375, "rewards/accuracy_reward/std": 0.23509246110916138, "step": 2683 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1985.0, "completions/mean_length": 1643.130859375, "completions/mean_terminated_length": 628.1849365234375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9162754971409064, "frac_reward_zero_std": 0.75, "grad_norm": 45.78015294625495, "learning_rate": 1.1934282081018023e-07, "loss": 0.0138, "num_tokens": 2536626189.0, "reward": 0.08203125, "reward_std": 0.09655161201953888, "rewards/accuracy_reward/mean": 0.08467742055654526, "rewards/accuracy_reward/std": 0.278682142496109, "step": 2684 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.724609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1867.0, "completions/mean_length": 1635.498046875, "completions/mean_terminated_length": 550.1205444335938, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9166168814542972, "frac_reward_zero_std": 0.75, "grad_norm": 62.569540792616586, "learning_rate": 1.1918763624325942e-07, "loss": 0.0275, "num_tokens": 2537536332.0, "reward": 0.03515625, "reward_std": 0.08493967354297638, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 2685 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.732421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2019.0, "completions/mean_length": 1686.67578125, "completions/mean_terminated_length": 697.649658203125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.916958265767688, "frac_reward_zero_std": 0.71875, "grad_norm": 39.352672277183366, "learning_rate": 1.1903306313569242e-07, "loss": 0.0289, "num_tokens": 2538470694.0, "reward": 0.04296875, "reward_std": 0.09891509264707565, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 2686 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1959.0, "completions/mean_length": 1576.2265625, "completions/mean_terminated_length": 610.2142944335938, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9172996500810787, "frac_reward_zero_std": 0.71875, "grad_norm": 7.483814036372228, "learning_rate": 1.1887910170686726e-07, "loss": 0.0262, "num_tokens": 2539344250.0, "reward": 0.05078125, "reward_std": 0.09099511802196503, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 2687 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.689453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2009.0, "completions/mean_length": 1625.06640625, "completions/mean_terminated_length": 686.1005859375, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.9176410343944695, "frac_reward_zero_std": 0.8125, "grad_norm": 66.21176219708461, "learning_rate": 1.1872575217530422e-07, "loss": 0.0042, "num_tokens": 2540249292.0, "reward": 0.0234375, "reward_std": 0.05974511057138443, "rewards/accuracy_reward/mean": 0.0234375, "rewards/accuracy_reward/std": 0.15143637359142303, "step": 2688 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1848.0, "completions/mean_length": 1614.158203125, "completions/mean_terminated_length": 624.1090087890625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.9179824187078603, "frac_reward_zero_std": 0.78125, "grad_norm": 16.858899246772374, "learning_rate": 1.1857301475865477e-07, "loss": 0.0273, "num_tokens": 2541145277.0, "reward": 0.041015625, "reward_std": 0.07780591398477554, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 2689 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.740234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1875.0, "completions/mean_length": 1677.619140625, "completions/mean_terminated_length": 622.1729736328125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9183238030212512, "frac_reward_zero_std": 0.625, "grad_norm": 40.27477226381563, "learning_rate": 1.1842088967370173e-07, "loss": 0.0084, "num_tokens": 2542088234.0, "reward": 0.07421875, "reward_std": 0.13915522396564484, "rewards/accuracy_reward/mean": 0.07421875, "rewards/accuracy_reward/std": 0.2623828947544098, "step": 2690 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.70703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1817.0, "completions/mean_length": 1629.306640625, "completions/mean_terminated_length": 618.8599853515625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.918665187334642, "frac_reward_zero_std": 0.78125, "grad_norm": 6.943651406759533, "learning_rate": 1.1826937713635902e-07, "loss": 0.017, "num_tokens": 2542993783.0, "reward": 0.04296875, "reward_std": 0.0843411535024643, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 2691 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.755859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1964.0, "completions/mean_length": 1687.072265625, "completions/mean_terminated_length": 569.6400146484375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9190065716480328, "frac_reward_zero_std": 0.65625, "grad_norm": 14.010187967142512, "learning_rate": 1.1811847736167078e-07, "loss": 0.035, "num_tokens": 2543928236.0, "reward": 0.091796875, "reward_std": 0.11879178136587143, "rewards/accuracy_reward/mean": 0.091796875, "rewards/accuracy_reward/std": 0.289021372795105, "step": 2692 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2026.0, "completions/mean_length": 1705.23828125, "completions/mean_terminated_length": 585.550048828125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9193479559614236, "frac_reward_zero_std": 0.75, "grad_norm": 153.15785904263205, "learning_rate": 1.1796819056381175e-07, "loss": 0.0141, "num_tokens": 2544882630.0, "reward": 0.0703125, "reward_std": 0.10229447484016418, "rewards/accuracy_reward/mean": 0.0703125, "rewards/accuracy_reward/std": 0.25592297315597534, "step": 2693 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.810546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2044.0, "completions/mean_length": 1780.873046875, "completions/mean_terminated_length": 638.6288452148438, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9196893402748144, "frac_reward_zero_std": 0.8125, "grad_norm": 89.2986222257337, "learning_rate": 1.178185169560865e-07, "loss": 0.002, "num_tokens": 2545875605.0, "reward": 0.0546875, "reward_std": 0.07873040437698364, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 2694 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.724609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1958.0, "completions/mean_length": 1649.33984375, "completions/mean_terminated_length": 600.3829345703125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9200307245882051, "frac_reward_zero_std": 0.65625, "grad_norm": 30.40658796290671, "learning_rate": 1.1766945675092938e-07, "loss": 0.0157, "num_tokens": 2546789107.0, "reward": 0.072265625, "reward_std": 0.12556564807891846, "rewards/accuracy_reward/mean": 0.072265625, "rewards/accuracy_reward/std": 0.2591804563999176, "step": 2695 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1993.0, "completions/mean_length": 1674.5, "completions/mean_terminated_length": 682.05712890625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9203721089015959, "frac_reward_zero_std": 0.8125, "grad_norm": 13.427209547897768, "learning_rate": 1.1752101015990404e-07, "loss": 0.0133, "num_tokens": 2547727987.0, "reward": 0.033203125, "reward_std": 0.06602106243371964, "rewards/accuracy_reward/mean": 0.033203125, "rewards/accuracy_reward/std": 0.17934183776378632, "step": 2696 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1987.0, "completions/mean_length": 1666.92578125, "completions/mean_terminated_length": 654.6500244140625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.9207134932149867, "frac_reward_zero_std": 0.8125, "grad_norm": 39.34895516214777, "learning_rate": 1.1737317739370323e-07, "loss": 0.0118, "num_tokens": 2548663261.0, "reward": 0.048828125, "reward_std": 0.0775209367275238, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 2697 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.697265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1992.0, "completions/mean_length": 1614.271484375, "completions/mean_terminated_length": 615.2967529296875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9210548775283776, "frac_reward_zero_std": 0.8125, "grad_norm": 95.13349357725528, "learning_rate": 1.172259586621487e-07, "loss": 0.0087, "num_tokens": 2549563128.0, "reward": 0.056640625, "reward_std": 0.08309652656316757, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 2698 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.791015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1899.0, "completions/mean_length": 1729.455078125, "completions/mean_terminated_length": 523.7476196289062, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9213962618417684, "frac_reward_zero_std": 0.84375, "grad_norm": 10.563887941527359, "learning_rate": 1.170793541741903e-07, "loss": 0.0142, "num_tokens": 2550532513.0, "reward": 0.037109375, "reward_std": 0.06405794620513916, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 2699 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.720703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1739.0, "completions/mean_length": 1634.998046875, "completions/mean_terminated_length": 569.2797241210938, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9217376461551592, "frac_reward_zero_std": 0.78125, "grad_norm": 78.0721216098307, "learning_rate": 1.169333641379065e-07, "loss": -0.0029, "num_tokens": 2551449840.0, "reward": 0.021484375, "reward_std": 0.06519509106874466, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 2700 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1981.0, "completions/mean_length": 1681.302734375, "completions/mean_terminated_length": 667.49267578125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.92207903046855, "frac_reward_zero_std": 0.84375, "grad_norm": 24.034663805380703, "learning_rate": 1.167879887605032e-07, "loss": 0.0044, "num_tokens": 2552383979.0, "reward": 0.01171875, "reward_std": 0.04192390665411949, "rewards/accuracy_reward/mean": 0.01171875, "rewards/accuracy_reward/std": 0.10772226005792618, "step": 2701 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.697265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2022.0, "completions/mean_length": 1623.3359375, "completions/mean_terminated_length": 645.2387084960938, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9224204147819408, "frac_reward_zero_std": 0.6875, "grad_norm": 32.66605559967687, "learning_rate": 1.1664322824831437e-07, "loss": 0.0087, "num_tokens": 2553288887.0, "reward": 0.072265625, "reward_std": 0.12004335969686508, "rewards/accuracy_reward/mean": 0.072265625, "rewards/accuracy_reward/std": 0.2591804563999176, "step": 2702 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.705078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1721.0, "completions/mean_length": 1640.97265625, "completions/mean_terminated_length": 667.8807983398438, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9227617990953315, "frac_reward_zero_std": 0.71875, "grad_norm": 119.11291744511516, "learning_rate": 1.1649908280680094e-07, "loss": 0.0232, "num_tokens": 2554206729.0, "reward": 0.052734375, "reward_std": 0.10095052421092987, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 2703 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.712890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1872.0, "completions/mean_length": 1622.822265625, "completions/mean_terminated_length": 567.1088256835938, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9231031834087223, "frac_reward_zero_std": 0.78125, "grad_norm": 91.50595019641537, "learning_rate": 1.1635555264055105e-07, "loss": 0.0227, "num_tokens": 2555119742.0, "reward": 0.041015625, "reward_std": 0.08279120922088623, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 2704 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.693359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2042.0, "completions/mean_length": 1602.443359375, "completions/mean_terminated_length": 595.2993774414062, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9234445677221131, "frac_reward_zero_std": 0.84375, "grad_norm": 65.47751476389558, "learning_rate": 1.1621263795327965e-07, "loss": 0.0003, "num_tokens": 2556019793.0, "reward": 0.041015625, "reward_std": 0.06548243016004562, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 2705 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.712890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2018.0, "completions/mean_length": 1628.95703125, "completions/mean_terminated_length": 588.4761962890625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.923785952035504, "frac_reward_zero_std": 0.65625, "grad_norm": 50.20686603189005, "learning_rate": 1.1607033894782782e-07, "loss": 0.0047, "num_tokens": 2556931211.0, "reward": 0.095703125, "reward_std": 0.13593661785125732, "rewards/accuracy_reward/mean": 0.095703125, "rewards/accuracy_reward/std": 0.2944713830947876, "step": 2706 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.775390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2029.0, "completions/mean_length": 1724.6171875, "completions/mean_terminated_length": 608.2434692382812, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.9241273363488948, "frac_reward_zero_std": 0.59375, "grad_norm": 118.78194004543495, "learning_rate": 1.1592865582616306e-07, "loss": 0.0263, "num_tokens": 2557897047.0, "reward": 0.08203125, "reward_std": 0.15197205543518066, "rewards/accuracy_reward/mean": 0.08203125, "rewards/accuracy_reward/std": 0.2746807038784027, "step": 2707 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.66015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1892.0, "completions/mean_length": 1530.962890625, "completions/mean_terminated_length": 526.6034545898438, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9244687206622856, "frac_reward_zero_std": 0.75, "grad_norm": 29.48698281235695, "learning_rate": 1.1578758878937856e-07, "loss": -0.0057, "num_tokens": 2558754084.0, "reward": 0.033203125, "reward_std": 0.08489333093166351, "rewards/accuracy_reward/mean": 0.033203125, "rewards/accuracy_reward/std": 0.17934183776378632, "step": 2708 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1786.0, "completions/mean_length": 1657.443359375, "completions/mean_terminated_length": 619.6785888671875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.9248101049756764, "frac_reward_zero_std": 0.78125, "grad_norm": 24.305023217987692, "learning_rate": 1.1564713803769327e-07, "loss": -0.0014, "num_tokens": 2559685383.0, "reward": 0.060546875, "reward_std": 0.09226682782173157, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 2709 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.798828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2024.0, "completions/mean_length": 1770.923828125, "completions/mean_terminated_length": 671.2621459960938, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9251514892890672, "frac_reward_zero_std": 0.875, "grad_norm": 3.011637490709759, "learning_rate": 1.1550730377045126e-07, "loss": 0.0171, "num_tokens": 2560673504.0, "reward": 0.013671875, "reward_std": 0.03889618441462517, "rewards/accuracy_reward/mean": 0.013671875, "rewards/accuracy_reward/std": 0.1162383034825325, "step": 2710 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2041.0, "completions/mean_length": 1681.56640625, "completions/mean_terminated_length": 668.8235473632812, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.925492873602458, "frac_reward_zero_std": 0.84375, "grad_norm": 41.004865879357595, "learning_rate": 1.1536808618612175e-07, "loss": -0.0023, "num_tokens": 2561619506.0, "reward": 0.046875, "reward_std": 0.06739920377731323, "rewards/accuracy_reward/mean": 0.046875, "rewards/accuracy_reward/std": 0.21157780289649963, "step": 2711 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.72265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1998.0, "completions/mean_length": 1654.224609375, "completions/mean_terminated_length": 628.1901245117188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9258342579158487, "frac_reward_zero_std": 0.78125, "grad_norm": 84.52424197698116, "learning_rate": 1.1522948548229875e-07, "loss": 0.0394, "num_tokens": 2562541621.0, "reward": 0.08203125, "reward_std": 0.09637914597988129, "rewards/accuracy_reward/mean": 0.08203125, "rewards/accuracy_reward/std": 0.2746807038784027, "step": 2712 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.72265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1736.0, "completions/mean_length": 1647.44921875, "completions/mean_terminated_length": 603.7605590820312, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.9261756422292395, "frac_reward_zero_std": 0.71875, "grad_norm": 43.66863055563476, "learning_rate": 1.150915018557004e-07, "loss": 0.0072, "num_tokens": 2563460987.0, "reward": 0.0546875, "reward_std": 0.10716674476861954, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 2713 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1881.0, "completions/mean_length": 1623.728515625, "completions/mean_terminated_length": 560.1438598632812, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9265170265426304, "frac_reward_zero_std": 0.8125, "grad_norm": 25.863699183515383, "learning_rate": 1.1495413550216933e-07, "loss": 0.0164, "num_tokens": 2564367264.0, "reward": 0.02734375, "reward_std": 0.0635918527841568, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 2714 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.685546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1764.0, "completions/mean_length": 1576.21875, "completions/mean_terminated_length": 547.677001953125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9268584108560212, "frac_reward_zero_std": 0.78125, "grad_norm": 36.15677394177097, "learning_rate": 1.1481738661667192e-07, "loss": 0.006, "num_tokens": 2565250128.0, "reward": 0.05078125, "reward_std": 0.08599072694778442, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 2715 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.681640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2020.0, "completions/mean_length": 1627.0390625, "completions/mean_terminated_length": 725.7177734375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.927199795169412, "frac_reward_zero_std": 0.78125, "grad_norm": 20.22435238516492, "learning_rate": 1.1468125539329826e-07, "loss": 0.0117, "num_tokens": 2566163012.0, "reward": 0.04296875, "reward_std": 0.08368149399757385, "rewards/accuracy_reward/mean": 0.04435483738780022, "rewards/accuracy_reward/std": 0.2060900777578354, "step": 2716 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1839.0, "completions/mean_length": 1683.9921875, "completions/mean_terminated_length": 716.7714233398438, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9275411794828028, "frac_reward_zero_std": 0.75, "grad_norm": 10.40211046814643, "learning_rate": 1.1454574202526165e-07, "loss": 0.0081, "num_tokens": 2567096784.0, "reward": 0.04296875, "reward_std": 0.08274346590042114, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 2717 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.82421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2035.0, "completions/mean_length": 1826.90234375, "completions/mean_terminated_length": 790.822265625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.9278825637961936, "frac_reward_zero_std": 0.78125, "grad_norm": 57.33836108659242, "learning_rate": 1.1441084670489857e-07, "loss": -0.0051, "num_tokens": 2568117566.0, "reward": 0.046875, "reward_std": 0.08296996355056763, "rewards/accuracy_reward/mean": 0.046875, "rewards/accuracy_reward/std": 0.21157780289649963, "step": 2718 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1913.0, "completions/mean_length": 1655.236328125, "completions/mean_terminated_length": 651.5069580078125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9282239481095844, "frac_reward_zero_std": 0.8125, "grad_norm": 29.284215973339307, "learning_rate": 1.1427656962366829e-07, "loss": -0.0046, "num_tokens": 2569043095.0, "reward": 0.044921875, "reward_std": 0.06409074366092682, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 2719 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.68359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1874.0, "completions/mean_length": 1585.908203125, "completions/mean_terminated_length": 587.5617065429688, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.9285653324229751, "frac_reward_zero_std": 0.78125, "grad_norm": 17.209251690673963, "learning_rate": 1.1414291097215244e-07, "loss": 0.0119, "num_tokens": 2569926408.0, "reward": 0.037109375, "reward_std": 0.08042868226766586, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 2720 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.724609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1982.0, "completions/mean_length": 1668.431640625, "completions/mean_terminated_length": 669.709228515625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9289067167363659, "frac_reward_zero_std": 0.75, "grad_norm": 23.160842792222304, "learning_rate": 1.1400987094005518e-07, "loss": 0.0048, "num_tokens": 2570858789.0, "reward": 0.05859375, "reward_std": 0.09259280562400818, "rewards/accuracy_reward/mean": 0.05859375, "rewards/accuracy_reward/std": 0.23509246110916138, "step": 2721 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.748046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1950.0, "completions/mean_length": 1694.875, "completions/mean_terminated_length": 646.4495849609375, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.9292481010497567, "frac_reward_zero_std": 0.6875, "grad_norm": 241.11601149971844, "learning_rate": 1.1387744971620236e-07, "loss": 0.0039, "num_tokens": 2571804261.0, "reward": 0.1015625, "reward_std": 0.12356442213058472, "rewards/accuracy_reward/mean": 0.1015625, "rewards/accuracy_reward/std": 0.30236753821372986, "step": 2722 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.681640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1819.0, "completions/mean_length": 1568.29296875, "completions/mean_terminated_length": 541.190185546875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9295894853631476, "frac_reward_zero_std": 0.75, "grad_norm": 168.52586375181815, "learning_rate": 1.137456474885418e-07, "loss": 0.0182, "num_tokens": 2572685355.0, "reward": 0.0625, "reward_std": 0.103780098259449, "rewards/accuracy_reward/mean": 0.06451612710952759, "rewards/accuracy_reward/std": 0.2459181249141693, "step": 2723 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.705078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1929.0, "completions/mean_length": 1623.955078125, "completions/mean_terminated_length": 610.1788330078125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9299308696765384, "frac_reward_zero_std": 0.75, "grad_norm": 48.45278978914982, "learning_rate": 1.136144644441426e-07, "loss": 0.0194, "num_tokens": 2573592212.0, "reward": 0.03515625, "reward_std": 0.0868166983127594, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 2724 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.681640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1847.0, "completions/mean_length": 1581.89453125, "completions/mean_terminated_length": 584.2637939453125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9302722539899292, "frac_reward_zero_std": 0.6875, "grad_norm": 144.97910134563475, "learning_rate": 1.1348390076919519e-07, "loss": 0.0038, "num_tokens": 2574489374.0, "reward": 0.06640625, "reward_std": 0.11117979884147644, "rewards/accuracy_reward/mean": 0.06640625, "rewards/accuracy_reward/std": 0.2492343932390213, "step": 2725 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2016.0, "completions/mean_length": 1737.857421875, "completions/mean_terminated_length": 724.7250366210938, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.93061363830332, "frac_reward_zero_std": 0.75, "grad_norm": 92.00130666992754, "learning_rate": 1.1335395664901071e-07, "loss": 0.0037, "num_tokens": 2575465221.0, "reward": 0.044921875, "reward_std": 0.09001073986291885, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 2726 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.720703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1959.0, "completions/mean_length": 1663.509765625, "completions/mean_terminated_length": 671.3636474609375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9309550226167108, "frac_reward_zero_std": 0.75, "grad_norm": 5.960058399792734, "learning_rate": 1.1322463226802109e-07, "loss": 0.0272, "num_tokens": 2576397754.0, "reward": 0.05078125, "reward_std": 0.0908287987112999, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 2727 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1882.0, "completions/mean_length": 1644.53125, "completions/mean_terminated_length": 652.2162475585938, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.9312964069301015, "frac_reward_zero_std": 0.78125, "grad_norm": 28.097684850926672, "learning_rate": 1.1309592780977867e-07, "loss": 0.0218, "num_tokens": 2577317818.0, "reward": 0.05078125, "reward_std": 0.08171837776899338, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 2728 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.67578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1975.0, "completions/mean_length": 1589.921875, "completions/mean_terminated_length": 635.1325073242188, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.9316377912434923, "frac_reward_zero_std": 0.75, "grad_norm": 80.1537414749105, "learning_rate": 1.1296784345695585e-07, "loss": 0.0139, "num_tokens": 2578212338.0, "reward": 0.05859375, "reward_std": 0.09959384053945541, "rewards/accuracy_reward/mean": 0.05859375, "rewards/accuracy_reward/std": 0.23509246110916138, "step": 2729 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.767578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1960.0, "completions/mean_length": 1710.052734375, "completions/mean_terminated_length": 593.9747924804688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9319791755568831, "frac_reward_zero_std": 0.75, "grad_norm": 55.22754725480669, "learning_rate": 1.1284037939134502e-07, "loss": 0.0165, "num_tokens": 2579164253.0, "reward": 0.0390625, "reward_std": 0.08698301762342453, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 2730 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1878.0, "completions/mean_length": 1673.841796875, "completions/mean_terminated_length": 735.883544921875, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.932320559870274, "frac_reward_zero_std": 0.78125, "grad_norm": 30.091282368687303, "learning_rate": 1.1271353579385804e-07, "loss": 0.0093, "num_tokens": 2580096716.0, "reward": 0.03515625, "reward_std": 0.06728477776050568, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 2731 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.701171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1934.0, "completions/mean_length": 1616.3984375, "completions/mean_terminated_length": 603.686279296875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9326619441836648, "frac_reward_zero_std": 0.71875, "grad_norm": 15.01036058263, "learning_rate": 1.1258731284452616e-07, "loss": 0.0219, "num_tokens": 2581000264.0, "reward": 0.060546875, "reward_std": 0.10051832348108292, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 2732 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1987.0, "completions/mean_length": 1662.4453125, "completions/mean_terminated_length": 695.9177856445312, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9330033284970556, "frac_reward_zero_std": 0.65625, "grad_norm": 157.00944534543342, "learning_rate": 1.1246171072249991e-07, "loss": 0.0334, "num_tokens": 2581922508.0, "reward": 0.05078125, "reward_std": 0.11399346590042114, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 2733 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.732421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2048.0, "completions/mean_length": 1672.53125, "completions/mean_terminated_length": 644.788330078125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9333447128104464, "frac_reward_zero_std": 0.75, "grad_norm": 67.66819588666246, "learning_rate": 1.1233672960604836e-07, "loss": 0.0053, "num_tokens": 2582863100.0, "reward": 0.083984375, "reward_std": 0.09488959610462189, "rewards/accuracy_reward/mean": 0.083984375, "rewards/accuracy_reward/std": 0.2776356339454651, "step": 2734 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.638671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2028.0, "completions/mean_length": 1573.80859375, "completions/mean_terminated_length": 735.6432495117188, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.9336860971238372, "frac_reward_zero_std": 0.75, "grad_norm": 37.98364460807045, "learning_rate": 1.1221236967255949e-07, "loss": 0.0158, "num_tokens": 2583743226.0, "reward": 0.048828125, "reward_std": 0.09995920211076736, "rewards/accuracy_reward/mean": 0.05040322616696358, "rewards/accuracy_reward/std": 0.21899642050266266, "step": 2735 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2009.0, "completions/mean_length": 1534.20703125, "completions/mean_terminated_length": 618.3152465820312, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9340274814372279, "frac_reward_zero_std": 0.78125, "grad_norm": 44.67205170004858, "learning_rate": 1.1208863109853938e-07, "loss": 0.0095, "num_tokens": 2584603140.0, "reward": 0.087890625, "reward_std": 0.09291848540306091, "rewards/accuracy_reward/mean": 0.087890625, "rewards/accuracy_reward/std": 0.2834126651287079, "step": 2736 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.70703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1823.0, "completions/mean_length": 1633.912109375, "completions/mean_terminated_length": 634.5800170898438, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9343688657506187, "frac_reward_zero_std": 0.71875, "grad_norm": 115.5863804021599, "learning_rate": 1.1196551405961232e-07, "loss": 0.0076, "num_tokens": 2585514087.0, "reward": 0.078125, "reward_std": 0.11599055677652359, "rewards/accuracy_reward/mean": 0.08064515888690948, "rewards/accuracy_reward/std": 0.2725643217563629, "step": 2737 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.697265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1851.0, "completions/mean_length": 1628.193359375, "completions/mean_terminated_length": 661.5806274414062, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.9347102500640095, "frac_reward_zero_std": 0.71875, "grad_norm": 20.030953089130673, "learning_rate": 1.1184301873052052e-07, "loss": 0.0247, "num_tokens": 2586430154.0, "reward": 0.056640625, "reward_std": 0.09793071448802948, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 2738 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.716796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1984.0, "completions/mean_length": 1656.51171875, "completions/mean_terminated_length": 665.641357421875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.9350516343774004, "frac_reward_zero_std": 0.78125, "grad_norm": 15.98536389419781, "learning_rate": 1.1172114528512358e-07, "loss": 0.0124, "num_tokens": 2587352576.0, "reward": 0.02734375, "reward_std": 0.07041998207569122, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 2739 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1778.0, "completions/mean_length": 1700.65625, "completions/mean_terminated_length": 566.0, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.9353930186907912, "frac_reward_zero_std": 0.84375, "grad_norm": 66.11732070506618, "learning_rate": 1.115998938963986e-07, "loss": -0.0109, "num_tokens": 2588304880.0, "reward": 0.025390625, "reward_std": 0.05913964658975601, "rewards/accuracy_reward/mean": 0.025390625, "rewards/accuracy_reward/std": 0.15746226906776428, "step": 2740 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.669921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1880.0, "completions/mean_length": 1563.783203125, "completions/mean_terminated_length": 581.023681640625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.935734403004182, "frac_reward_zero_std": 0.8125, "grad_norm": 51.30875452422642, "learning_rate": 1.1147926473643973e-07, "loss": 0.018, "num_tokens": 2589178081.0, "reward": 0.076171875, "reward_std": 0.06713200360536575, "rewards/accuracy_reward/mean": 0.076171875, "rewards/accuracy_reward/std": 0.26553234457969666, "step": 2741 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1756.0, "completions/mean_length": 1648.302734375, "completions/mean_terminated_length": 626.8541870117188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9360757873175728, "frac_reward_zero_std": 0.71875, "grad_norm": 12.031535529426744, "learning_rate": 1.1135925797645812e-07, "loss": 0.0281, "num_tokens": 2590099516.0, "reward": 0.05078125, "reward_std": 0.10107050836086273, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 2742 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.669921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2003.0, "completions/mean_length": 1601.560546875, "completions/mean_terminated_length": 695.473388671875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9364171716309636, "frac_reward_zero_std": 0.65625, "grad_norm": 91.9273160364728, "learning_rate": 1.1123987378678127e-07, "loss": -0.0145, "num_tokens": 2591001099.0, "reward": 0.06640625, "reward_std": 0.12862756848335266, "rewards/accuracy_reward/mean": 0.06640625, "rewards/accuracy_reward/std": 0.2492343932390213, "step": 2743 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.69140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2028.0, "completions/mean_length": 1624.4921875, "completions/mean_terminated_length": 675.6203002929688, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.9367585559443543, "frac_reward_zero_std": 0.6875, "grad_norm": 58.1854467136591, "learning_rate": 1.1112111233685323e-07, "loss": 0.0133, "num_tokens": 2591915447.0, "reward": 0.080078125, "reward_std": 0.11256343126296997, "rewards/accuracy_reward/mean": 0.080078125, "rewards/accuracy_reward/std": 0.271679550409317, "step": 2744 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.681640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1945.0, "completions/mean_length": 1611.607421875, "completions/mean_terminated_length": 677.245361328125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9370999402577451, "frac_reward_zero_std": 0.6875, "grad_norm": 52.80460932369611, "learning_rate": 1.1100297379523423e-07, "loss": 0.0325, "num_tokens": 2592813342.0, "reward": 0.078125, "reward_std": 0.12452814728021622, "rewards/accuracy_reward/mean": 0.078125, "rewards/accuracy_reward/std": 0.26863065361976624, "step": 2745 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.732421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2027.0, "completions/mean_length": 1676.80078125, "completions/mean_terminated_length": 660.7445068359375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9374413245711359, "frac_reward_zero_std": 0.6875, "grad_norm": 186.4548638799342, "learning_rate": 1.108854583296002e-07, "loss": 0.0135, "num_tokens": 2593764888.0, "reward": 0.03515625, "reward_std": 0.09358368813991547, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 2746 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.736328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1923.0, "completions/mean_length": 1677.2578125, "completions/mean_terminated_length": 642.34814453125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9377827088845268, "frac_reward_zero_std": 0.875, "grad_norm": 23.75898137694367, "learning_rate": 1.1076856610674298e-07, "loss": 0.0023, "num_tokens": 2594706412.0, "reward": 0.041015625, "reward_std": 0.051707521080970764, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 2747 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.607421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1986.0, "completions/mean_length": 1497.279296875, "completions/mean_terminated_length": 645.1691284179688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9381240931979176, "frac_reward_zero_std": 0.6875, "grad_norm": 35.61007948697366, "learning_rate": 1.106522972925696e-07, "loss": 0.0227, "num_tokens": 2595543211.0, "reward": 0.056640625, "reward_std": 0.11520528793334961, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 2748 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.65234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1995.0, "completions/mean_length": 1548.552734375, "completions/mean_terminated_length": 611.3876342773438, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9384654775113084, "frac_reward_zero_std": 0.78125, "grad_norm": 44.124940602045285, "learning_rate": 1.1053665205210249e-07, "loss": 0.006, "num_tokens": 2596410790.0, "reward": 0.044921875, "reward_std": 0.08626452088356018, "rewards/accuracy_reward/mean": 0.04791666567325592, "rewards/accuracy_reward/std": 0.21381278336048126, "step": 2749 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1888.0, "completions/mean_length": 1543.865234375, "completions/mean_terminated_length": 614.0166625976562, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9388068618246992, "frac_reward_zero_std": 0.75, "grad_norm": 44.82521235730442, "learning_rate": 1.1042163054947881e-07, "loss": 0.008, "num_tokens": 2597278337.0, "reward": 0.0546875, "reward_std": 0.09908141195774078, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 2750 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.72265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1955.0, "completions/mean_length": 1641.416015625, "completions/mean_terminated_length": 582.0070190429688, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.93914824613809, "frac_reward_zero_std": 0.90625, "grad_norm": 51.29181778064116, "learning_rate": 1.1030723294795055e-07, "loss": -0.0011, "num_tokens": 2598198502.0, "reward": 0.021484375, "reward_std": 0.036420635879039764, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 2751 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.74609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2031.0, "completions/mean_length": 1679.267578125, "completions/mean_terminated_length": 595.7615356445312, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9394896304514807, "frac_reward_zero_std": 0.8125, "grad_norm": 75.63200972463699, "learning_rate": 1.1019345940988427e-07, "loss": 0.0057, "num_tokens": 2599148095.0, "reward": 0.03125, "reward_std": 0.06805649399757385, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 2752 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.708984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2028.0, "completions/mean_length": 1632.703125, "completions/mean_terminated_length": 621.2684326171875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9398310147648715, "frac_reward_zero_std": 0.75, "grad_norm": 22.784256299772256, "learning_rate": 1.1008031009676061e-07, "loss": 0.007, "num_tokens": 2600065271.0, "reward": 0.107421875, "reward_std": 0.10524433106184006, "rewards/accuracy_reward/mean": 0.107421875, "rewards/accuracy_reward/std": 0.30995169281959534, "step": 2753 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.697265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1972.0, "completions/mean_length": 1608.4765625, "completions/mean_terminated_length": 596.1548461914062, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9401723990782623, "frac_reward_zero_std": 0.6875, "grad_norm": 21.234588876096534, "learning_rate": 1.0996778516917438e-07, "loss": 0.0025, "num_tokens": 2600965051.0, "reward": 0.048828125, "reward_std": 0.1072334349155426, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 2754 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1924.0, "completions/mean_length": 1612.96484375, "completions/mean_terminated_length": 620.1923217773438, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9405137833916531, "frac_reward_zero_std": 0.6875, "grad_norm": 132.7547833821959, "learning_rate": 1.0985588478683407e-07, "loss": -0.0065, "num_tokens": 2601869161.0, "reward": 0.068359375, "reward_std": 0.11258251965045929, "rewards/accuracy_reward/mean": 0.068359375, "rewards/accuracy_reward/std": 0.25260838866233826, "step": 2755 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.677734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2038.0, "completions/mean_length": 1600.240234375, "completions/mean_terminated_length": 658.9030151367188, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.940855167705044, "frac_reward_zero_std": 0.59375, "grad_norm": 112.81075642327981, "learning_rate": 1.0974460910856182e-07, "loss": 0.0141, "num_tokens": 2602767652.0, "reward": 0.15234375, "reward_std": 0.17551058530807495, "rewards/accuracy_reward/mean": 0.15234375, "rewards/accuracy_reward/std": 0.35970520973205566, "step": 2756 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.736328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1977.0, "completions/mean_length": 1686.90625, "completions/mean_terminated_length": 678.5184936523438, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9411965520184348, "frac_reward_zero_std": 0.71875, "grad_norm": 38.14059005238338, "learning_rate": 1.0963395829229322e-07, "loss": 0.0071, "num_tokens": 2603712660.0, "reward": 0.08203125, "reward_std": 0.11199754476547241, "rewards/accuracy_reward/mean": 0.08203125, "rewards/accuracy_reward/std": 0.2746807038784027, "step": 2757 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.66796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1734.0, "completions/mean_length": 1574.728515625, "completions/mean_terminated_length": 622.61767578125, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.9415379363318256, "frac_reward_zero_std": 0.875, "grad_norm": 18.127502583790335, "learning_rate": 1.0952393249507669e-07, "loss": 0.0043, "num_tokens": 2604598345.0, "reward": 0.029296875, "reward_std": 0.04859926179051399, "rewards/accuracy_reward/mean": 0.030241934582591057, "rewards/accuracy_reward/std": 0.1714252382516861, "step": 2758 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1983.0, "completions/mean_length": 1727.6328125, "completions/mean_terminated_length": 634.336181640625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9418793206452164, "frac_reward_zero_std": 0.84375, "grad_norm": 33.65955123591645, "learning_rate": 1.0941453187307386e-07, "loss": 0.0072, "num_tokens": 2605575837.0, "reward": 0.02734375, "reward_std": 0.0580955371260643, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 2759 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.634765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1959.0, "completions/mean_length": 1534.009765625, "completions/mean_terminated_length": 640.7112426757812, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9422207049586071, "frac_reward_zero_std": 0.78125, "grad_norm": 60.070594072096036, "learning_rate": 1.0930575658155882e-07, "loss": -0.0078, "num_tokens": 2606437602.0, "reward": 0.044921875, "reward_std": 0.08384227007627487, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 2760 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.712890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2019.0, "completions/mean_length": 1645.064453125, "completions/mean_terminated_length": 644.5782470703125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9425620892719979, "frac_reward_zero_std": 0.84375, "grad_norm": 80.31843034298011, "learning_rate": 1.0919760677491827e-07, "loss": -0.0115, "num_tokens": 2607369187.0, "reward": 0.02734375, "reward_std": 0.05311024188995361, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 2761 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.712890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1992.0, "completions/mean_length": 1657.345703125, "completions/mean_terminated_length": 687.3536987304688, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9429034735853887, "frac_reward_zero_std": 0.65625, "grad_norm": 55.031603826544206, "learning_rate": 1.0909008260665102e-07, "loss": 0.0086, "num_tokens": 2608292164.0, "reward": 0.076171875, "reward_std": 0.1269710510969162, "rewards/accuracy_reward/mean": 0.076171875, "rewards/accuracy_reward/std": 0.26553234457969666, "step": 2762 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.689453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2025.0, "completions/mean_length": 1624.078125, "completions/mean_terminated_length": 682.918212890625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.9432448578987795, "frac_reward_zero_std": 0.71875, "grad_norm": 20.479561532627358, "learning_rate": 1.0898318422936796e-07, "loss": 0.0234, "num_tokens": 2609202188.0, "reward": 0.046875, "reward_std": 0.09668608754873276, "rewards/accuracy_reward/mean": 0.046875, "rewards/accuracy_reward/std": 0.21157780289649963, "step": 2763 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.689453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1947.0, "completions/mean_length": 1618.396484375, "completions/mean_terminated_length": 664.6226196289062, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.9435862422121704, "frac_reward_zero_std": 0.625, "grad_norm": 33.901773629948465, "learning_rate": 1.0887691179479182e-07, "loss": 0.02, "num_tokens": 2610106903.0, "reward": 0.09375, "reward_std": 0.12933912873268127, "rewards/accuracy_reward/mean": 0.09375, "rewards/accuracy_reward/std": 0.29176566004753113, "step": 2764 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.69140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1977.0, "completions/mean_length": 1615.375, "completions/mean_terminated_length": 646.0759887695312, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9439276265255612, "frac_reward_zero_std": 0.78125, "grad_norm": 53.82307433033624, "learning_rate": 1.0877126545375688e-07, "loss": -0.0011, "num_tokens": 2611007559.0, "reward": 0.03125, "reward_std": 0.07152432948350906, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 2765 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2016.0, "completions/mean_length": 1592.484375, "completions/mean_terminated_length": 625.9024047851562, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.944269010838952, "frac_reward_zero_std": 0.71875, "grad_norm": 20.78402405448913, "learning_rate": 1.0866624535620878e-07, "loss": 0.0274, "num_tokens": 2611894287.0, "reward": 0.072265625, "reward_std": 0.10606793314218521, "rewards/accuracy_reward/mean": 0.072265625, "rewards/accuracy_reward/std": 0.2591804563999176, "step": 2766 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.689453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1948.0, "completions/mean_length": 1618.310546875, "completions/mean_terminated_length": 664.3458862304688, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.9446103951523428, "frac_reward_zero_std": 0.75, "grad_norm": 41.76218707746861, "learning_rate": 1.0856185165120433e-07, "loss": 0.0131, "num_tokens": 2612793422.0, "reward": 0.0546875, "reward_std": 0.09737758338451385, "rewards/accuracy_reward/mean": 0.05645161122083664, "rewards/accuracy_reward/std": 0.23102475702762604, "step": 2767 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.70703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1984.0, "completions/mean_length": 1627.8671875, "completions/mean_terminated_length": 613.9466552734375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9449517794657335, "frac_reward_zero_std": 0.75, "grad_norm": 74.51461093563086, "learning_rate": 1.0845808448691141e-07, "loss": 0.0193, "num_tokens": 2613704890.0, "reward": 0.064453125, "reward_std": 0.09804278612136841, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 2768 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.673828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2002.0, "completions/mean_length": 1617.619140625, "completions/mean_terminated_length": 728.509033203125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9452931637791243, "frac_reward_zero_std": 0.6875, "grad_norm": 17.538055564272263, "learning_rate": 1.0835494401060835e-07, "loss": 0.0308, "num_tokens": 2614605127.0, "reward": 0.060546875, "reward_std": 0.11081992089748383, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 2769 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2037.0, "completions/mean_length": 1627.767578125, "completions/mean_terminated_length": 632.4802856445312, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.9456345480925151, "frac_reward_zero_std": 0.84375, "grad_norm": 28.037376991573883, "learning_rate": 1.0825243036868424e-07, "loss": -0.0077, "num_tokens": 2615524032.0, "reward": 0.03125, "reward_std": 0.054907046258449554, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 2770 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.673828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2018.0, "completions/mean_length": 1565.544921875, "completions/mean_terminated_length": 568.8563232421875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9459759324059059, "frac_reward_zero_std": 0.71875, "grad_norm": 59.25682114292323, "learning_rate": 1.081505437066386e-07, "loss": 0.0131, "num_tokens": 2616403655.0, "reward": 0.052734375, "reward_std": 0.09875431656837463, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 2771 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.73828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2007.0, "completions/mean_length": 1678.3671875, "completions/mean_terminated_length": 635.671630859375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9463173167192968, "frac_reward_zero_std": 0.75, "grad_norm": 13.608687420642724, "learning_rate": 1.0804928416908073e-07, "loss": 0.0141, "num_tokens": 2617352851.0, "reward": 0.025390625, "reward_std": 0.07344770431518555, "rewards/accuracy_reward/mean": 0.025390625, "rewards/accuracy_reward/std": 0.15746226906776428, "step": 2772 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2043.0, "completions/mean_length": 1658.947265625, "completions/mean_terminated_length": 625.1785888671875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9466587010326876, "frac_reward_zero_std": 0.6875, "grad_norm": 26.398674378618324, "learning_rate": 1.0794865189973011e-07, "loss": 0.0031, "num_tokens": 2618280008.0, "reward": 0.083984375, "reward_std": 0.10854348540306091, "rewards/accuracy_reward/mean": 0.083984375, "rewards/accuracy_reward/std": 0.2776356339454651, "step": 2773 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.740234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1924.0, "completions/mean_length": 1701.197265625, "completions/mean_terminated_length": 713.26318359375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9470000853460784, "frac_reward_zero_std": 0.8125, "grad_norm": 14.626394930538885, "learning_rate": 1.0784864704141585e-07, "loss": 0.0121, "num_tokens": 2619228205.0, "reward": 0.046875, "reward_std": 0.07104447484016418, "rewards/accuracy_reward/mean": 0.046875, "rewards/accuracy_reward/std": 0.21157780289649963, "step": 2774 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1659.0, "completions/mean_length": 1552.5234375, "completions/mean_terminated_length": 537.9761962890625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.9473414696594692, "frac_reward_zero_std": 0.84375, "grad_norm": 6.401036062291176, "learning_rate": 1.0774926973607648e-07, "loss": 0.0223, "num_tokens": 2620094409.0, "reward": 0.03515625, "reward_std": 0.058320626616477966, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 2775 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.720703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1944.0, "completions/mean_length": 1624.53125, "completions/mean_terminated_length": 531.80419921875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.9476828539728599, "frac_reward_zero_std": 0.59375, "grad_norm": 76.81720096755522, "learning_rate": 1.0765052012475998e-07, "loss": 0.0102, "num_tokens": 2621001897.0, "reward": 0.115234375, "reward_std": 0.15497520565986633, "rewards/accuracy_reward/mean": 0.115234375, "rewards/accuracy_reward/std": 0.3196168541908264, "step": 2776 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.744140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1771.0, "completions/mean_length": 1684.212890625, "completions/mean_terminated_length": 626.1755981445312, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9480242382862507, "frac_reward_zero_std": 0.78125, "grad_norm": 18.185943554605814, "learning_rate": 1.0755239834762326e-07, "loss": -0.0038, "num_tokens": 2621946630.0, "reward": 0.02734375, "reward_std": 0.07284127175807953, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 2777 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.65625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1757.0, "completions/mean_length": 1544.9453125, "completions/mean_terminated_length": 584.5681762695312, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9483656225996415, "frac_reward_zero_std": 0.71875, "grad_norm": 40.33436865191403, "learning_rate": 1.0745490454393239e-07, "loss": 0.0179, "num_tokens": 2622817530.0, "reward": 0.04296875, "reward_std": 0.09820214658975601, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 2778 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1633.0, "completions/mean_length": 1633.5703125, "completions/mean_terminated_length": 574.4722290039062, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9487070069130323, "frac_reward_zero_std": 0.71875, "grad_norm": 23.7167455906824, "learning_rate": 1.0735803885206191e-07, "loss": 0.0191, "num_tokens": 2623740334.0, "reward": 0.0546875, "reward_std": 0.10810574144124985, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 2779 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.681640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1945.0, "completions/mean_length": 1597.3125, "completions/mean_terminated_length": 632.3435668945312, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9490483912264231, "frac_reward_zero_std": 0.78125, "grad_norm": 31.43848003449339, "learning_rate": 1.0726180140949497e-07, "loss": 0.0202, "num_tokens": 2624638670.0, "reward": 0.0625, "reward_std": 0.09055041521787643, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.2422981858253479, "step": 2780 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.626953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1934.0, "completions/mean_length": 1524.224609375, "completions/mean_terminated_length": 643.952880859375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.949389775539814, "frac_reward_zero_std": 0.65625, "grad_norm": 298.9405124362075, "learning_rate": 1.0716619235282295e-07, "loss": -0.0019, "num_tokens": 2625502833.0, "reward": 0.078125, "reward_std": 0.13090261816978455, "rewards/accuracy_reward/mean": 0.078125, "rewards/accuracy_reward/std": 0.26863065361976624, "step": 2781 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.716796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1763.0, "completions/mean_length": 1657.978515625, "completions/mean_terminated_length": 670.8206787109375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9497311598532048, "frac_reward_zero_std": 0.875, "grad_norm": 32.508942537629025, "learning_rate": 1.0707121181774556e-07, "loss": -0.001, "num_tokens": 2626425366.0, "reward": 0.041015625, "reward_std": 0.052558064460754395, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 2782 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.775390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1869.0, "completions/mean_length": 1731.576171875, "completions/mean_terminated_length": 639.22607421875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.9500725441665956, "frac_reward_zero_std": 0.78125, "grad_norm": 43.158924354935245, "learning_rate": 1.0697685993907009e-07, "loss": 0.005, "num_tokens": 2627382765.0, "reward": 0.076171875, "reward_std": 0.09994550049304962, "rewards/accuracy_reward/mean": 0.076171875, "rewards/accuracy_reward/std": 0.26553234457969666, "step": 2783 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.72265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1791.0, "completions/mean_length": 1669.041015625, "completions/mean_terminated_length": 681.6126708984375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9504139284799863, "frac_reward_zero_std": 0.6875, "grad_norm": 22.36082329580779, "learning_rate": 1.0688313685071194e-07, "loss": 0.0558, "num_tokens": 2628306546.0, "reward": 0.08203125, "reward_std": 0.11379435658454895, "rewards/accuracy_reward/mean": 0.08203125, "rewards/accuracy_reward/std": 0.2746807038784027, "step": 2784 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.712890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1875.0, "completions/mean_length": 1636.912109375, "completions/mean_terminated_length": 616.1836547851562, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9507553127933771, "frac_reward_zero_std": 0.84375, "grad_norm": 2.4984989202737102, "learning_rate": 1.067900426856939e-07, "loss": 0.012, "num_tokens": 2629230789.0, "reward": 0.041015625, "reward_std": 0.04957009106874466, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 2785 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.669921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1894.0, "completions/mean_length": 1557.091796875, "completions/mean_terminated_length": 560.75146484375, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.9510966971067679, "frac_reward_zero_std": 0.71875, "grad_norm": 87.08025709085688, "learning_rate": 1.0669757757614602e-07, "loss": 0.0063, "num_tokens": 2630104660.0, "reward": 0.072265625, "reward_std": 0.11190377175807953, "rewards/accuracy_reward/mean": 0.072265625, "rewards/accuracy_reward/std": 0.2591804563999176, "step": 2786 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.708984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2033.0, "completions/mean_length": 1647.5703125, "completions/mean_terminated_length": 672.02685546875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.9514380814201587, "frac_reward_zero_std": 0.65625, "grad_norm": 57.59161770725123, "learning_rate": 1.0660574165330567e-07, "loss": 0.0135, "num_tokens": 2631026280.0, "reward": 0.07421875, "reward_std": 0.13241459429264069, "rewards/accuracy_reward/mean": 0.07421875, "rewards/accuracy_reward/std": 0.2623828947544098, "step": 2787 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.67578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1872.0, "completions/mean_length": 1591.00390625, "completions/mean_terminated_length": 638.4698486328125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9517794657335495, "frac_reward_zero_std": 0.8125, "grad_norm": 5.165057838133421, "learning_rate": 1.065145350475171e-07, "loss": 0.0196, "num_tokens": 2631920746.0, "reward": 0.048828125, "reward_std": 0.069866843521595, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 2788 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.759765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1994.0, "completions/mean_length": 1699.037109375, "completions/mean_terminated_length": 595.406494140625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9521208500469404, "frac_reward_zero_std": 0.84375, "grad_norm": 17.10636160723882, "learning_rate": 1.0642395788823144e-07, "loss": 0.0076, "num_tokens": 2632869517.0, "reward": 0.029296875, "reward_std": 0.055033616721630096, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 2789 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.740234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1997.0, "completions/mean_length": 1679.794921875, "completions/mean_terminated_length": 630.5488891601562, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9524622343603312, "frac_reward_zero_std": 0.6875, "grad_norm": 57.983493154565714, "learning_rate": 1.0633401030400637e-07, "loss": 0.0042, "num_tokens": 2633816116.0, "reward": 0.056640625, "reward_std": 0.1038198471069336, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 2790 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1928.0, "completions/mean_length": 1711.263671875, "completions/mean_terminated_length": 611.2583618164062, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.952803618673722, "frac_reward_zero_std": 0.6875, "grad_norm": 17.62248292375537, "learning_rate": 1.0624469242250607e-07, "loss": 0.0217, "num_tokens": 2634773547.0, "reward": 0.072265625, "reward_std": 0.1204371303319931, "rewards/accuracy_reward/mean": 0.072265625, "rewards/accuracy_reward/std": 0.2591804563999176, "step": 2791 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1801.0, "completions/mean_length": 1613.625, "completions/mean_terminated_length": 658.0, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9531450029871127, "frac_reward_zero_std": 0.71875, "grad_norm": 45.22998985733543, "learning_rate": 1.0615600437050094e-07, "loss": 0.0015, "num_tokens": 2635678219.0, "reward": 0.064453125, "reward_std": 0.08995649218559265, "rewards/accuracy_reward/mean": 0.06653226166963577, "rewards/accuracy_reward/std": 0.2494617998600006, "step": 2792 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.69140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1981.0, "completions/mean_length": 1606.55078125, "completions/mean_terminated_length": 617.4810180664062, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9534863873005035, "frac_reward_zero_std": 0.625, "grad_norm": 41.57475349412186, "learning_rate": 1.0606794627386739e-07, "loss": 0.0239, "num_tokens": 2636578533.0, "reward": 0.064453125, "reward_std": 0.13596117496490479, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 2793 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.74609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2024.0, "completions/mean_length": 1679.35546875, "completions/mean_terminated_length": 596.107666015625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9538277716138943, "frac_reward_zero_std": 0.875, "grad_norm": 339.1737721130537, "learning_rate": 1.0598051825758785e-07, "loss": 0.0078, "num_tokens": 2637517531.0, "reward": 0.021484375, "reward_std": 0.04478531330823898, "rewards/accuracy_reward/mean": 0.02217741869390011, "rewards/accuracy_reward/std": 0.14740893244743347, "step": 2794 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.75390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1979.0, "completions/mean_length": 1714.236328125, "completions/mean_terminated_length": 691.7540283203125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9541691559272851, "frac_reward_zero_std": 0.78125, "grad_norm": 76.41940991299751, "learning_rate": 1.0589372044575035e-07, "loss": 0.0046, "num_tokens": 2638473316.0, "reward": 0.05859375, "reward_std": 0.09644614160060883, "rewards/accuracy_reward/mean": 0.05859375, "rewards/accuracy_reward/std": 0.23509246110916138, "step": 2795 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1866.0, "completions/mean_length": 1696.296875, "completions/mean_terminated_length": 683.8181762695312, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.9545105402406759, "frac_reward_zero_std": 0.71875, "grad_norm": 67.88778799587612, "learning_rate": 1.0580755296154857e-07, "loss": 0.0097, "num_tokens": 2639414524.0, "reward": 0.052734375, "reward_std": 0.10453138500452042, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 2796 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.77734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2037.0, "completions/mean_length": 1745.779296875, "completions/mean_terminated_length": 691.0877075195312, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9548519245540668, "frac_reward_zero_std": 0.6875, "grad_norm": 76.90895384380907, "learning_rate": 1.0572201592728136e-07, "loss": 0.0061, "num_tokens": 2640393947.0, "reward": 0.041015625, "reward_std": 0.09820909798145294, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 2797 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.748046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1973.0, "completions/mean_length": 1686.509765625, "completions/mean_terminated_length": 613.248046875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.9551933088674576, "frac_reward_zero_std": 0.96875, "grad_norm": 34.13231580675639, "learning_rate": 1.0563710946435309e-07, "loss": -0.001, "num_tokens": 2641329760.0, "reward": 0.009765625, "reward_std": 0.014959799125790596, "rewards/accuracy_reward/mean": 0.009765625, "rewards/accuracy_reward/std": 0.09843364357948303, "step": 2798 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.66796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1964.0, "completions/mean_length": 1562.044921875, "completions/mean_terminated_length": 584.4176635742188, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9555346931808484, "frac_reward_zero_std": 0.71875, "grad_norm": 62.972893025440385, "learning_rate": 1.0555283369327283e-07, "loss": 0.0196, "num_tokens": 2642204839.0, "reward": 0.099609375, "reward_std": 0.11646249890327454, "rewards/accuracy_reward/mean": 0.10282257944345474, "rewards/accuracy_reward/std": 0.30403366684913635, "step": 2799 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.732421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1889.0, "completions/mean_length": 1669.064453125, "completions/mean_terminated_length": 631.8320922851562, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.9558760774942391, "frac_reward_zero_std": 0.8125, "grad_norm": 4.8162017538011455, "learning_rate": 1.0546918873365457e-07, "loss": 0.0113, "num_tokens": 2643133128.0, "reward": 0.029296875, "reward_std": 0.0688755214214325, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 2800 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.701171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1849.0, "completions/mean_length": 1629.677734375, "completions/mean_terminated_length": 648.1242065429688, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9562174618076299, "frac_reward_zero_std": 0.8125, "grad_norm": 50.36986136012113, "learning_rate": 1.0538617470421715e-07, "loss": 0.0169, "num_tokens": 2644050003.0, "reward": 0.072265625, "reward_std": 0.08329563587903976, "rewards/accuracy_reward/mean": 0.072265625, "rewards/accuracy_reward/std": 0.2591804563999176, "step": 2801 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.740234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1887.0, "completions/mean_length": 1659.3515625, "completions/mean_terminated_length": 551.849609375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9565588461210207, "frac_reward_zero_std": 0.65625, "grad_norm": 15.521473220634554, "learning_rate": 1.0530379172278375e-07, "loss": 0.0134, "num_tokens": 2644976567.0, "reward": 0.078125, "reward_std": 0.11983070522546768, "rewards/accuracy_reward/mean": 0.078125, "rewards/accuracy_reward/std": 0.26863065361976624, "step": 2802 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.79296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1750.0, "completions/mean_length": 1780.99609375, "completions/mean_terminated_length": 758.3207397460938, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9569002304344115, "frac_reward_zero_std": 0.6875, "grad_norm": 54.45079795062958, "learning_rate": 1.0522203990628196e-07, "loss": 0.0122, "num_tokens": 2645965877.0, "reward": 0.064453125, "reward_std": 0.11409997940063477, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 2803 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1841.0, "completions/mean_length": 1647.619140625, "completions/mean_terminated_length": 662.898681640625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9572416147478023, "frac_reward_zero_std": 0.71875, "grad_norm": 30.585366899428866, "learning_rate": 1.0514091937074349e-07, "loss": 0.0075, "num_tokens": 2646888594.0, "reward": 0.087890625, "reward_std": 0.11674422025680542, "rewards/accuracy_reward/mean": 0.087890625, "rewards/accuracy_reward/std": 0.2834126651287079, "step": 2804 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2032.0, "completions/mean_length": 1699.06640625, "completions/mean_terminated_length": 559.2166748046875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9575829990611932, "frac_reward_zero_std": 0.8125, "grad_norm": 21.93363708169347, "learning_rate": 1.050604302313042e-07, "loss": -0.0009, "num_tokens": 2647835060.0, "reward": 0.021484375, "reward_std": 0.05688370764255524, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 2805 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.76953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2024.0, "completions/mean_length": 1752.119140625, "completions/mean_terminated_length": 764.177978515625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.957924383374584, "frac_reward_zero_std": 0.6875, "grad_norm": 23.87044937716521, "learning_rate": 1.0498057260220361e-07, "loss": 0.0219, "num_tokens": 2648811681.0, "reward": 0.0390625, "reward_std": 0.1009584367275238, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 2806 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1870.0, "completions/mean_length": 1627.259765625, "completions/mean_terminated_length": 592.4662475585938, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.9582657676879748, "frac_reward_zero_std": 0.8125, "grad_norm": 5.046464069521894, "learning_rate": 1.0490134659678501e-07, "loss": 0.0012, "num_tokens": 2649722934.0, "reward": 0.05078125, "reward_std": 0.07261618971824646, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 2807 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.779296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2024.0, "completions/mean_length": 1746.93359375, "completions/mean_terminated_length": 683.8760986328125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9586071520013656, "frac_reward_zero_std": 0.8125, "grad_norm": 33.981074276649544, "learning_rate": 1.0482275232749527e-07, "loss": -0.0007, "num_tokens": 2650698244.0, "reward": 0.060546875, "reward_std": 0.06570751965045929, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 2808 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.705078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1960.0, "completions/mean_length": 1603.669921875, "completions/mean_terminated_length": 541.3973388671875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9589485363147563, "frac_reward_zero_std": 0.6875, "grad_norm": 63.151196722593745, "learning_rate": 1.0474478990588456e-07, "loss": 0.0097, "num_tokens": 2651597083.0, "reward": 0.048828125, "reward_std": 0.10310593992471695, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 2809 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.697265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2015.0, "completions/mean_length": 1641.763671875, "completions/mean_terminated_length": 706.1096801757812, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9592899206281471, "frac_reward_zero_std": 0.8125, "grad_norm": 34.988260941838924, "learning_rate": 1.0466745944260631e-07, "loss": -0.0037, "num_tokens": 2652516258.0, "reward": 0.0234375, "reward_std": 0.05974511429667473, "rewards/accuracy_reward/mean": 0.0234375, "rewards/accuracy_reward/std": 0.15143637359142303, "step": 2810 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.759765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2016.0, "completions/mean_length": 1717.16796875, "completions/mean_terminated_length": 671.3495483398438, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.9596313049415379, "frac_reward_zero_std": 0.65625, "grad_norm": 55.5623160185692, "learning_rate": 1.0459076104741699e-07, "loss": 0.0234, "num_tokens": 2653483592.0, "reward": 0.078125, "reward_std": 0.13841770589351654, "rewards/accuracy_reward/mean": 0.078125, "rewards/accuracy_reward/std": 0.26863065361976624, "step": 2811 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.724609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2017.0, "completions/mean_length": 1661.779296875, "completions/mean_terminated_length": 645.5531616210938, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9599726892549287, "frac_reward_zero_std": 0.78125, "grad_norm": 59.45413373747896, "learning_rate": 1.045146948291758e-07, "loss": -0.0079, "num_tokens": 2654414679.0, "reward": 0.02734375, "reward_std": 0.07190324366092682, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 2812 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.732421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1912.0, "completions/mean_length": 1646.609375, "completions/mean_terminated_length": 548.2554931640625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9603140735683195, "frac_reward_zero_std": 0.65625, "grad_norm": 40.86011064635546, "learning_rate": 1.0443926089584498e-07, "loss": 0.0155, "num_tokens": 2655343391.0, "reward": 0.060546875, "reward_std": 0.12559986114501953, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 2813 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.736328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1993.0, "completions/mean_length": 1676.404296875, "completions/mean_terminated_length": 638.6889038085938, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.9606554578817104, "frac_reward_zero_std": 0.65625, "grad_norm": 55.44845658479631, "learning_rate": 1.0436445935448916e-07, "loss": 0.0357, "num_tokens": 2656277534.0, "reward": 0.078125, "reward_std": 0.12275754660367966, "rewards/accuracy_reward/mean": 0.078125, "rewards/accuracy_reward/std": 0.26863065361976624, "step": 2814 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.73828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1964.0, "completions/mean_length": 1670.787109375, "completions/mean_terminated_length": 606.7089233398438, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9609968421951012, "frac_reward_zero_std": 0.875, "grad_norm": 90.74118891929331, "learning_rate": 1.0429029031127539e-07, "loss": 0.014, "num_tokens": 2657212945.0, "reward": 0.017578125, "reward_std": 0.04125870764255524, "rewards/accuracy_reward/mean": 0.017578125, "rewards/accuracy_reward/std": 0.13154059648513794, "step": 2815 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.69921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1979.0, "completions/mean_length": 1628.625, "completions/mean_terminated_length": 653.7142944335938, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.961338226508492, "frac_reward_zero_std": 0.78125, "grad_norm": 18.735630472699523, "learning_rate": 1.042167538714731e-07, "loss": 0.0027, "num_tokens": 2658126657.0, "reward": 0.037109375, "reward_std": 0.07509727776050568, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 2816 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2041.0, "completions/mean_length": 1614.47265625, "completions/mean_terminated_length": 694.5487670898438, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.9616796108218827, "frac_reward_zero_std": 0.75, "grad_norm": 24.325912136683133, "learning_rate": 1.0414385013945384e-07, "loss": 0.02, "num_tokens": 2659026147.0, "reward": 0.07421875, "reward_std": 0.09765692055225372, "rewards/accuracy_reward/mean": 0.07421875, "rewards/accuracy_reward/std": 0.2623828947544098, "step": 2817 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.654296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1925.0, "completions/mean_length": 1567.1796875, "completions/mean_terminated_length": 657.1525268554688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9620209951352735, "frac_reward_zero_std": 0.875, "grad_norm": 17.69442269370351, "learning_rate": 1.040715792186911e-07, "loss": 0.0101, "num_tokens": 2659902079.0, "reward": 0.02734375, "reward_std": 0.044233135879039764, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 2818 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.736328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1766.0, "completions/mean_length": 1677.353515625, "completions/mean_terminated_length": 642.2888793945312, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.9623623794486643, "frac_reward_zero_std": 0.71875, "grad_norm": 38.01990825802955, "learning_rate": 1.0399994121176025e-07, "loss": 0.0056, "num_tokens": 2660832932.0, "reward": 0.041015625, "reward_std": 0.07933683693408966, "rewards/accuracy_reward/mean": 0.04233871027827263, "rewards/accuracy_reward/std": 0.2015640139579773, "step": 2819 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.685546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1786.0, "completions/mean_length": 1625.408203125, "completions/mean_terminated_length": 704.1055908203125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9627037637620551, "frac_reward_zero_std": 0.71875, "grad_norm": 46.61658634146071, "learning_rate": 1.039289362203383e-07, "loss": 0.0158, "num_tokens": 2661735397.0, "reward": 0.060546875, "reward_std": 0.10766562819480896, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 2820 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.728515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1823.0, "completions/mean_length": 1654.796875, "completions/mean_terminated_length": 599.6546630859375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.963045148075446, "frac_reward_zero_std": 0.71875, "grad_norm": 69.15783122166368, "learning_rate": 1.0385856434520387e-07, "loss": 0.0164, "num_tokens": 2662655485.0, "reward": 0.08984375, "reward_std": 0.10579650849103928, "rewards/accuracy_reward/mean": 0.08984375, "rewards/accuracy_reward/std": 0.2862374484539032, "step": 2821 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1972.0, "completions/mean_length": 1693.587890625, "completions/mean_terminated_length": 673.3106079101562, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9633865323888368, "frac_reward_zero_std": 0.65625, "grad_norm": 28.06453506220154, "learning_rate": 1.0378882568623697e-07, "loss": 0.0261, "num_tokens": 2663607258.0, "reward": 0.111328125, "reward_std": 0.14288339018821716, "rewards/accuracy_reward/mean": 0.111328125, "rewards/accuracy_reward/std": 0.31484565138816833, "step": 2822 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.72265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1855.0, "completions/mean_length": 1636.26953125, "completions/mean_terminated_length": 563.45068359375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9637279167022276, "frac_reward_zero_std": 0.78125, "grad_norm": 18.20562581942902, "learning_rate": 1.037197203424189e-07, "loss": -0.0064, "num_tokens": 2664519076.0, "reward": 0.03125, "reward_std": 0.06574726104736328, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 2823 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.693359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1998.0, "completions/mean_length": 1630.193359375, "completions/mean_terminated_length": 685.8344116210938, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.9640693010156184, "frac_reward_zero_std": 0.75, "grad_norm": 29.577812066301465, "learning_rate": 1.0365124841183199e-07, "loss": 0.0065, "num_tokens": 2665438599.0, "reward": 0.060546875, "reward_std": 0.09110259264707565, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 2824 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.69921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1950.0, "completions/mean_length": 1617.326171875, "completions/mean_terminated_length": 616.1493530273438, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9644106853290091, "frac_reward_zero_std": 0.75, "grad_norm": 33.911037396901534, "learning_rate": 1.0358340999165966e-07, "loss": -0.0029, "num_tokens": 2666347822.0, "reward": 0.064453125, "reward_std": 0.08879335969686508, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 2825 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1937.0, "completions/mean_length": 1627.984375, "completions/mean_terminated_length": 669.4871826171875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.9647520696423999, "frac_reward_zero_std": 0.75, "grad_norm": 118.65605462662212, "learning_rate": 1.035162051781861e-07, "loss": 0.0135, "num_tokens": 2667262870.0, "reward": 0.048828125, "reward_std": 0.09958265721797943, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 2826 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.65625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2014.0, "completions/mean_length": 1564.994140625, "completions/mean_terminated_length": 642.89208984375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9650934539557907, "frac_reward_zero_std": 0.8125, "grad_norm": 5.879935933961238, "learning_rate": 1.0344963406679633e-07, "loss": 0.0255, "num_tokens": 2668136819.0, "reward": 0.037109375, "reward_std": 0.07399433106184006, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 2827 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.70703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2011.0, "completions/mean_length": 1621.35546875, "completions/mean_terminated_length": 592.0933227539062, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9654348382691815, "frac_reward_zero_std": 0.8125, "grad_norm": 7.246639897713062, "learning_rate": 1.0338369675197584e-07, "loss": -0.0122, "num_tokens": 2669057049.0, "reward": 0.029296875, "reward_std": 0.06183479726314545, "rewards/accuracy_reward/mean": 0.029296875, "rewards/accuracy_reward/std": 0.16880230605602264, "step": 2828 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1920.0, "completions/mean_length": 1643.900390625, "completions/mean_terminated_length": 611.201416015625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9657762225825723, "frac_reward_zero_std": 0.625, "grad_norm": 183.4541280317545, "learning_rate": 1.0331839332731053e-07, "loss": 0.04, "num_tokens": 2669978902.0, "reward": 0.083984375, "reward_std": 0.14557647705078125, "rewards/accuracy_reward/mean": 0.083984375, "rewards/accuracy_reward/std": 0.2776356339454651, "step": 2829 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.666015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2011.0, "completions/mean_length": 1579.33984375, "completions/mean_terminated_length": 644.76025390625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9661176068959632, "frac_reward_zero_std": 0.65625, "grad_norm": 33.91036134659535, "learning_rate": 1.0325372388548673e-07, "loss": 0.0272, "num_tokens": 2670868756.0, "reward": 0.0859375, "reward_std": 0.1156454086303711, "rewards/accuracy_reward/mean": 0.0859375, "rewards/accuracy_reward/std": 0.28054583072662354, "step": 2830 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.74609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1679.0, "completions/mean_length": 1678.3359375, "completions/mean_terminated_length": 592.09228515625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.966458991209354, "frac_reward_zero_std": 0.75, "grad_norm": 37.96084090286949, "learning_rate": 1.0318968851829084e-07, "loss": 0.0173, "num_tokens": 2671804016.0, "reward": 0.064453125, "reward_std": 0.09656515717506409, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 2831 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.783203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1925.0, "completions/mean_length": 1751.162109375, "completions/mean_terminated_length": 678.8018188476562, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.9668003755227448, "frac_reward_zero_std": 0.78125, "grad_norm": 13.438861638100496, "learning_rate": 1.031262873166094e-07, "loss": 0.0102, "num_tokens": 2672778499.0, "reward": 0.072265625, "reward_std": 0.09212085604667664, "rewards/accuracy_reward/mean": 0.072265625, "rewards/accuracy_reward/std": 0.2591804563999176, "step": 2832 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.728515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1986.0, "completions/mean_length": 1674.5703125, "completions/mean_terminated_length": 672.4891967773438, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9671417598361355, "frac_reward_zero_std": 0.84375, "grad_norm": 43.35877301799867, "learning_rate": 1.0306352037042878e-07, "loss": 0.0009, "num_tokens": 2673705383.0, "reward": 0.0390625, "reward_std": 0.06310540437698364, "rewards/accuracy_reward/mean": 0.0390625, "rewards/accuracy_reward/std": 0.1939331740140915, "step": 2833 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.720703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2034.0, "completions/mean_length": 1661.263671875, "completions/mean_terminated_length": 663.3216552734375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9674831441495263, "frac_reward_zero_std": 0.75, "grad_norm": 23.77442352253742, "learning_rate": 1.0300138776883521e-07, "loss": 0.0075, "num_tokens": 2674631822.0, "reward": 0.02734375, "reward_std": 0.07685433328151703, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 2834 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1990.0, "completions/mean_length": 1640.572265625, "completions/mean_terminated_length": 619.2123413085938, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9678245284629171, "frac_reward_zero_std": 0.90625, "grad_norm": 5.553063967294221, "learning_rate": 1.0293988960001453e-07, "loss": -0.0018, "num_tokens": 2675539651.0, "reward": 0.005859375, "reward_std": 0.0234375, "rewards/accuracy_reward/mean": 0.005859375, "rewards/accuracy_reward/std": 0.07639661431312561, "step": 2835 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1851.0, "completions/mean_length": 1564.513671875, "completions/mean_terminated_length": 702.9619750976562, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.9681659127763079, "frac_reward_zero_std": 0.65625, "grad_norm": 99.31695707004648, "learning_rate": 1.0287902595125212e-07, "loss": 0.0176, "num_tokens": 2676420522.0, "reward": 0.0703125, "reward_std": 0.12999781966209412, "rewards/accuracy_reward/mean": 0.0703125, "rewards/accuracy_reward/std": 0.25592297315597534, "step": 2836 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.666015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1995.0, "completions/mean_length": 1602.0703125, "completions/mean_terminated_length": 712.8187255859375, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.9685072970896987, "frac_reward_zero_std": 0.75, "grad_norm": 39.407667955363884, "learning_rate": 1.0281879690893287e-07, "loss": 0.0022, "num_tokens": 2677317454.0, "reward": 0.048828125, "reward_std": 0.07797222584486008, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 2837 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.681640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2018.0, "completions/mean_length": 1606.453125, "completions/mean_terminated_length": 661.05517578125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9688486814030896, "frac_reward_zero_std": 0.8125, "grad_norm": 62.54421755793352, "learning_rate": 1.0275920255854082e-07, "loss": 0.0156, "num_tokens": 2678222534.0, "reward": 0.0625, "reward_std": 0.06409768760204315, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.2422981858253479, "step": 2838 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2023.0, "completions/mean_length": 1552.720703125, "completions/mean_terminated_length": 573.6802368164062, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9691900657164804, "frac_reward_zero_std": 0.71875, "grad_norm": 7.044876609113754, "learning_rate": 1.0270024298465929e-07, "loss": 0.0111, "num_tokens": 2679093863.0, "reward": 0.08203125, "reward_std": 0.11379435658454895, "rewards/accuracy_reward/mean": 0.08203125, "rewards/accuracy_reward/std": 0.2746807038784027, "step": 2839 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.666015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1943.0, "completions/mean_length": 1591.505859375, "completions/mean_terminated_length": 681.1871337890625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.9695314500298712, "frac_reward_zero_std": 0.875, "grad_norm": 28.719995002620692, "learning_rate": 1.0264191827097057e-07, "loss": 0.0081, "num_tokens": 2679991450.0, "reward": 0.015625, "reward_std": 0.04175759106874466, "rewards/accuracy_reward/mean": 0.015625, "rewards/accuracy_reward/std": 0.12414088100194931, "step": 2840 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.697265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2007.0, "completions/mean_length": 1642.142578125, "completions/mean_terminated_length": 707.3612670898438, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9698728343432619, "frac_reward_zero_std": 0.71875, "grad_norm": 7.560760169506948, "learning_rate": 1.0258422850025601e-07, "loss": 0.0043, "num_tokens": 2680913187.0, "reward": 0.041015625, "reward_std": 0.09511469304561615, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 2841 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.673828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2041.0, "completions/mean_length": 1590.009765625, "completions/mean_terminated_length": 644.1377563476562, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9702142186566527, "frac_reward_zero_std": 0.75, "grad_norm": 99.34078843576712, "learning_rate": 1.025271737543956e-07, "loss": 0.0103, "num_tokens": 2681813336.0, "reward": 0.037109375, "reward_std": 0.08725681155920029, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 2842 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2038.0, "completions/mean_length": 1624.525390625, "completions/mean_terminated_length": 658.4807739257812, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.9705556029700435, "frac_reward_zero_std": 0.71875, "grad_norm": 84.22504371034195, "learning_rate": 1.0247075411436815e-07, "loss": 0.0181, "num_tokens": 2682720581.0, "reward": 0.064453125, "reward_std": 0.11563889682292938, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 2843 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.650390625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1936.0, "completions/mean_length": 1583.681640625, "completions/mean_terminated_length": 719.8938598632812, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9708969872834343, "frac_reward_zero_std": 0.78125, "grad_norm": 71.82667001143582, "learning_rate": 1.0241496966025103e-07, "loss": 0.0184, "num_tokens": 2683611554.0, "reward": 0.0625, "reward_std": 0.09145425260066986, "rewards/accuracy_reward/mean": 0.0625, "rewards/accuracy_reward/std": 0.2422981858253479, "step": 2844 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.716796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2035.0, "completions/mean_length": 1624.330078125, "completions/mean_terminated_length": 552.0068969726562, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.9712383715968251, "frac_reward_zero_std": 0.875, "grad_norm": 5.761440234800375, "learning_rate": 1.0235982047121997e-07, "loss": -0.0012, "num_tokens": 2684524939.0, "reward": 0.044921875, "reward_std": 0.05287160724401474, "rewards/accuracy_reward/mean": 0.044921875, "rewards/accuracy_reward/std": 0.20733514428138733, "step": 2845 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.69921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1720.0, "completions/mean_length": 1604.26171875, "completions/mean_terminated_length": 572.7142944335938, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.971579755910216, "frac_reward_zero_std": 0.78125, "grad_norm": 31.834671505250128, "learning_rate": 1.0230530662554937e-07, "loss": 0.0049, "num_tokens": 2685424609.0, "reward": 0.064453125, "reward_std": 0.08241777867078781, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 2846 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.708984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1866.0, "completions/mean_length": 1651.60546875, "completions/mean_terminated_length": 685.8926391601562, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.9719211402236068, "frac_reward_zero_std": 0.75, "grad_norm": 17.713543753850242, "learning_rate": 1.0225142820061143e-07, "loss": -0.0032, "num_tokens": 2686348503.0, "reward": 0.09765625, "reward_std": 0.1158457025885582, "rewards/accuracy_reward/mean": 0.09765625, "rewards/accuracy_reward/std": 0.29713961482048035, "step": 2847 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.66796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1966.0, "completions/mean_length": 1576.96484375, "completions/mean_terminated_length": 629.3529663085938, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9722625245369976, "frac_reward_zero_std": 0.75, "grad_norm": 62.51068606352571, "learning_rate": 1.0219818527287691e-07, "loss": -0.0067, "num_tokens": 2687238277.0, "reward": 0.068359375, "reward_std": 0.09820909798145294, "rewards/accuracy_reward/mean": 0.068359375, "rewards/accuracy_reward/std": 0.25260838866233826, "step": 2848 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71484375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2044.0, "completions/mean_length": 1666.322265625, "completions/mean_terminated_length": 709.9383544921875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9726039088503883, "frac_reward_zero_std": 0.75, "grad_norm": 23.731152225323623, "learning_rate": 1.0214557791791425e-07, "loss": 0.005, "num_tokens": 2688177194.0, "reward": 0.02734375, "reward_std": 0.07630910724401474, "rewards/accuracy_reward/mean": 0.02822580561041832, "rewards/accuracy_reward/std": 0.1657845675945282, "step": 2849 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1979.0, "completions/mean_length": 1689.6875, "completions/mean_terminated_length": 699.058837890625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9729452931637791, "frac_reward_zero_std": 0.5625, "grad_norm": 49.90230468108034, "learning_rate": 1.0209360621039007e-07, "loss": 0.0196, "num_tokens": 2689120106.0, "reward": 0.1015625, "reward_std": 0.14755409955978394, "rewards/accuracy_reward/mean": 0.1015625, "rewards/accuracy_reward/std": 0.30236753821372986, "step": 2850 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.677734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1884.0, "completions/mean_length": 1590.05078125, "completions/mean_terminated_length": 627.5575561523438, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9732866774771699, "frac_reward_zero_std": 0.75, "grad_norm": 129.85617536703256, "learning_rate": 1.0204227022406866e-07, "loss": 0.022, "num_tokens": 2690023348.0, "reward": 0.09375, "reward_std": 0.10590954124927521, "rewards/accuracy_reward/mean": 0.09375, "rewards/accuracy_reward/std": 0.29176566004753113, "step": 2851 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.701171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1998.0, "completions/mean_length": 1625.298828125, "completions/mean_terminated_length": 633.7843017578125, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.9736280617905607, "frac_reward_zero_std": 0.8125, "grad_norm": 18.83981082691846, "learning_rate": 1.019915700318121e-07, "loss": 0.0007, "num_tokens": 2690941501.0, "reward": 0.041015625, "reward_std": 0.0706586092710495, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 2852 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.767578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1944.0, "completions/mean_length": 1725.42578125, "completions/mean_terminated_length": 660.11767578125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9739694461039515, "frac_reward_zero_std": 0.71875, "grad_norm": 40.20038317790714, "learning_rate": 1.0194150570558e-07, "loss": 0.004, "num_tokens": 2691906279.0, "reward": 0.05078125, "reward_std": 0.09225328266620636, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 2853 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2039.0, "completions/mean_length": 1673.64453125, "completions/mean_terminated_length": 752.9324340820312, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9743108304173423, "frac_reward_zero_std": 0.875, "grad_norm": 8.834654771958183, "learning_rate": 1.0189207731642956e-07, "loss": 0.0085, "num_tokens": 2692845809.0, "reward": 0.017578125, "reward_std": 0.04219770431518555, "rewards/accuracy_reward/mean": 0.017578125, "rewards/accuracy_reward/std": 0.13154059648513794, "step": 2854 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.634765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2007.0, "completions/mean_length": 1533.400390625, "completions/mean_terminated_length": 639.0427856445312, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9746522147307332, "frac_reward_zero_std": 0.65625, "grad_norm": 52.922226261703905, "learning_rate": 1.0184328493451527e-07, "loss": 0.0211, "num_tokens": 2693707918.0, "reward": 0.0546875, "reward_std": 0.11889266967773438, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 2855 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.732421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1797.0, "completions/mean_length": 1645.373046875, "completions/mean_terminated_length": 543.2919921875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.974993599044124, "frac_reward_zero_std": 0.78125, "grad_norm": 32.55534278262826, "learning_rate": 1.0179512862908905e-07, "loss": 0.0147, "num_tokens": 2694626237.0, "reward": 0.078125, "reward_std": 0.09554657340049744, "rewards/accuracy_reward/mean": 0.078125, "rewards/accuracy_reward/std": 0.26863065361976624, "step": 2856 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.74609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1950.0, "completions/mean_length": 1676.3046875, "completions/mean_terminated_length": 584.515380859375, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.9753349833575147, "frac_reward_zero_std": 0.78125, "grad_norm": 28.232388209437882, "learning_rate": 1.0174760846849994e-07, "loss": 0.0163, "num_tokens": 2695569193.0, "reward": 0.064453125, "reward_std": 0.08577118813991547, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 2857 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.62890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1968.0, "completions/mean_length": 1521.71484375, "completions/mean_terminated_length": 629.800048828125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9756763676709055, "frac_reward_zero_std": 0.6875, "grad_norm": 7.724777644549814, "learning_rate": 1.0170072452019414e-07, "loss": 0.0261, "num_tokens": 2696424263.0, "reward": 0.068359375, "reward_std": 0.11218452453613281, "rewards/accuracy_reward/mean": 0.068359375, "rewards/accuracy_reward/std": 0.25260838866233826, "step": 2858 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.72265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1938.0, "completions/mean_length": 1654.435546875, "completions/mean_terminated_length": 628.95068359375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.9760177519842963, "frac_reward_zero_std": 0.8125, "grad_norm": 40.930787500666476, "learning_rate": 1.0165447685071481e-07, "loss": 0.0022, "num_tokens": 2697355366.0, "reward": 0.03515625, "reward_std": 0.06987475603818893, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 2859 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2017.0, "completions/mean_length": 1673.90234375, "completions/mean_terminated_length": 787.881591796875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.9763591362976871, "frac_reward_zero_std": 0.78125, "grad_norm": 19.793387056519745, "learning_rate": 1.0160886552570211e-07, "loss": -0.0041, "num_tokens": 2698287332.0, "reward": 0.025390625, "reward_std": 0.06661957502365112, "rewards/accuracy_reward/mean": 0.025390625, "rewards/accuracy_reward/std": 0.15746226906776428, "step": 2860 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1975.0, "completions/mean_length": 1613.166015625, "completions/mean_terminated_length": 620.8526000976562, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9767005206110779, "frac_reward_zero_std": 0.78125, "grad_norm": 24.960496320120647, "learning_rate": 1.0156389060989289e-07, "loss": 0.0126, "num_tokens": 2699188601.0, "reward": 0.05078125, "reward_std": 0.06904878467321396, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 2861 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.69140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1998.0, "completions/mean_length": 1600.416015625, "completions/mean_terminated_length": 597.6012573242188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9770419049244687, "frac_reward_zero_std": 0.75, "grad_norm": 11.719228620061923, "learning_rate": 1.0151955216712089e-07, "loss": 0.0039, "num_tokens": 2700089150.0, "reward": 0.046875, "reward_std": 0.08698301762342453, "rewards/accuracy_reward/mean": 0.046875, "rewards/accuracy_reward/std": 0.21157780289649963, "step": 2862 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.62109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.0, "completions/mean_length": 1530.20703125, "completions/mean_terminated_length": 681.45361328125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9773832892378596, "frac_reward_zero_std": 0.84375, "grad_norm": 102.18074082610285, "learning_rate": 1.0147585026031632e-07, "loss": -0.001, "num_tokens": 2700950072.0, "reward": 0.03125, "reward_std": 0.05641176179051399, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 2863 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.708984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1950.0, "completions/mean_length": 1631.70703125, "completions/mean_terminated_length": 617.5167846679688, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9777246735512504, "frac_reward_zero_std": 0.6875, "grad_norm": 111.70927043383357, "learning_rate": 1.0143278495150619e-07, "loss": 0.0016, "num_tokens": 2701863762.0, "reward": 0.091796875, "reward_std": 0.1298992931842804, "rewards/accuracy_reward/mean": 0.091796875, "rewards/accuracy_reward/std": 0.289021372795105, "step": 2864 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.751953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2048.0, "completions/mean_length": 1671.275390625, "completions/mean_terminated_length": 529.2362060546875, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.9780660578646411, "frac_reward_zero_std": 0.71875, "grad_norm": 59.11837716668084, "learning_rate": 1.0139035630181373e-07, "loss": 0.0051, "num_tokens": 2702789039.0, "reward": 0.078125, "reward_std": 0.11575747281312943, "rewards/accuracy_reward/mean": 0.078125, "rewards/accuracy_reward/std": 0.26863065361976624, "step": 2865 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.736328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2012.0, "completions/mean_length": 1693.677734375, "completions/mean_terminated_length": 704.5481567382812, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.9784074421780319, "frac_reward_zero_std": 0.625, "grad_norm": 25.07626925278958, "learning_rate": 1.0134856437145871e-07, "loss": 0.0077, "num_tokens": 2703743274.0, "reward": 0.076171875, "reward_std": 0.13808965682983398, "rewards/accuracy_reward/mean": 0.0786290317773819, "rewards/accuracy_reward/std": 0.26943066716194153, "step": 2866 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.759765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1972.0, "completions/mean_length": 1721.595703125, "completions/mean_terminated_length": 689.3088989257812, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9787488264914227, "frac_reward_zero_std": 0.8125, "grad_norm": 34.63252352992463, "learning_rate": 1.0130740921975706e-07, "loss": 0.0136, "num_tokens": 2704700139.0, "reward": 0.048828125, "reward_std": 0.07686128467321396, "rewards/accuracy_reward/mean": 0.048828125, "rewards/accuracy_reward/std": 0.2157193273305893, "step": 2867 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.716796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1939.0, "completions/mean_length": 1649.58984375, "completions/mean_terminated_length": 641.6206665039062, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.9790902108048135, "frac_reward_zero_std": 0.65625, "grad_norm": 41.104247776451146, "learning_rate": 1.0126689090512102e-07, "loss": 0.0085, "num_tokens": 2705626697.0, "reward": 0.068359375, "reward_std": 0.12576615810394287, "rewards/accuracy_reward/mean": 0.07056451588869095, "rewards/accuracy_reward/std": 0.25635457038879395, "step": 2868 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.611328125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1895.0, "completions/mean_length": 1474.27734375, "completions/mean_terminated_length": 571.8894653320312, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9794315951182043, "frac_reward_zero_std": 0.625, "grad_norm": 122.05249741625943, "learning_rate": 1.0122700948505894e-07, "loss": 0.0627, "num_tokens": 2706455031.0, "reward": 0.080078125, "reward_std": 0.1427883505821228, "rewards/accuracy_reward/mean": 0.080078125, "rewards/accuracy_reward/std": 0.271679550409317, "step": 2869 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.73828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2015.0, "completions/mean_length": 1710.9375, "completions/mean_terminated_length": 760.119384765625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9797729794315951, "frac_reward_zero_std": 0.8125, "grad_norm": 21.736023448111542, "learning_rate": 1.0118776501617519e-07, "loss": 0.0045, "num_tokens": 2707418263.0, "reward": 0.0546875, "reward_std": 0.07355422526597977, "rewards/accuracy_reward/mean": 0.05645161122083664, "rewards/accuracy_reward/std": 0.23102474212646484, "step": 2870 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.712890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1793.0, "completions/mean_length": 1640.564453125, "completions/mean_terminated_length": 628.9047241210938, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.980114363744986, "frac_reward_zero_std": 0.8125, "grad_norm": 10.859671086388024, "learning_rate": 1.0114915755417014e-07, "loss": 0.0222, "num_tokens": 2708337496.0, "reward": 0.025390625, "reward_std": 0.06166848540306091, "rewards/accuracy_reward/mean": 0.025390625, "rewards/accuracy_reward/std": 0.15746226906776428, "step": 2871 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.701171875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1695.0, "completions/mean_length": 1609.236328125, "completions/mean_terminated_length": 579.7189331054688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9804557480583768, "frac_reward_zero_std": 0.75, "grad_norm": 13.69726999787243, "learning_rate": 1.0111118715383995e-07, "loss": 0.0017, "num_tokens": 2709238001.0, "reward": 0.072265625, "reward_std": 0.10556350648403168, "rewards/accuracy_reward/mean": 0.072265625, "rewards/accuracy_reward/std": 0.2591804563999176, "step": 2872 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.626953125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1975.0, "completions/mean_length": 1507.169921875, "completions/mean_terminated_length": 598.235595703125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.9807971323717675, "frac_reward_zero_std": 0.78125, "grad_norm": 53.87182218245254, "learning_rate": 1.0107385386907679e-07, "loss": 0.0013, "num_tokens": 2710084472.0, "reward": 0.076171875, "reward_std": 0.08588561415672302, "rewards/accuracy_reward/mean": 0.076171875, "rewards/accuracy_reward/std": 0.26553234457969666, "step": 2873 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.689453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1992.0, "completions/mean_length": 1614.236328125, "completions/mean_terminated_length": 651.2263793945312, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9811385166851583, "frac_reward_zero_std": 0.75, "grad_norm": 27.147053981870883, "learning_rate": 1.0103715775286826e-07, "loss": 0.0141, "num_tokens": 2710985505.0, "reward": 0.0859375, "reward_std": 0.10107050091028214, "rewards/accuracy_reward/mean": 0.0859375, "rewards/accuracy_reward/std": 0.28054583072662354, "step": 2874 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.748046875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1962.0, "completions/mean_length": 1667.822265625, "completions/mean_terminated_length": 539.43408203125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9814799009985491, "frac_reward_zero_std": 0.6875, "grad_norm": 37.13452790883539, "learning_rate": 1.010010988572979e-07, "loss": 0.021, "num_tokens": 2711925462.0, "reward": 0.076171875, "reward_std": 0.12066785991191864, "rewards/accuracy_reward/mean": 0.0786290317773819, "rewards/accuracy_reward/std": 0.26943066716194153, "step": 2875 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.681640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2039.0, "completions/mean_length": 1597.244140625, "completions/mean_terminated_length": 632.1288452148438, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.9818212853119399, "frac_reward_zero_std": 0.9375, "grad_norm": 4.168044770021395, "learning_rate": 1.0096567723354473e-07, "loss": 0.0146, "num_tokens": 2712817251.0, "reward": 0.01171875, "reward_std": 0.022772299125790596, "rewards/accuracy_reward/mean": 0.01171875, "rewards/accuracy_reward/std": 0.10772226005792618, "step": 2876 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.740234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1984.0, "completions/mean_length": 1661.76953125, "completions/mean_terminated_length": 561.1578979492188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9821626696253307, "frac_reward_zero_std": 0.78125, "grad_norm": 13.704344436381552, "learning_rate": 1.0093089293188319e-07, "loss": 0.0038, "num_tokens": 2713744157.0, "reward": 0.046875, "reward_std": 0.07662828266620636, "rewards/accuracy_reward/mean": 0.046875, "rewards/accuracy_reward/std": 0.21157780289649963, "step": 2877 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1836.0, "completions/mean_length": 1580.322265625, "completions/mean_terminated_length": 587.9329223632812, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9825040539387215, "frac_reward_zero_std": 0.71875, "grad_norm": 109.37150619209498, "learning_rate": 1.0089674600168329e-07, "loss": 0.032, "num_tokens": 2714629538.0, "reward": 0.041015625, "reward_std": 0.09511469304561615, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 2878 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2024.0, "completions/mean_length": 1584.43359375, "completions/mean_terminated_length": 668.0814208984375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9828454382521123, "frac_reward_zero_std": 0.6875, "grad_norm": 20.678783486882647, "learning_rate": 1.0086323649141032e-07, "loss": 0.0005, "num_tokens": 2715518400.0, "reward": 0.0703125, "reward_std": 0.11097268760204315, "rewards/accuracy_reward/mean": 0.0703125, "rewards/accuracy_reward/std": 0.25592297315597534, "step": 2879 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.689453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1868.0, "completions/mean_length": 1605.224609375, "completions/mean_terminated_length": 622.20751953125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.9831868225655032, "frac_reward_zero_std": 0.78125, "grad_norm": 14.445390615902841, "learning_rate": 1.0083036444862492e-07, "loss": 0.0205, "num_tokens": 2716408595.0, "reward": 0.052734375, "reward_std": 0.08791409432888031, "rewards/accuracy_reward/mean": 0.05443548411130905, "rewards/accuracy_reward/std": 0.227104052901268, "step": 2880 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.765625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2001.0, "completions/mean_length": 1701.2265625, "completions/mean_terminated_length": 568.433349609375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9835282068788939, "frac_reward_zero_std": 0.78125, "grad_norm": 24.624345397724333, "learning_rate": 1.0079812991998291e-07, "loss": 0.0282, "num_tokens": 2717361767.0, "reward": 0.0859375, "reward_std": 0.08792105317115784, "rewards/accuracy_reward/mean": 0.0859375, "rewards/accuracy_reward/std": 0.28054583072662354, "step": 2881 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.677734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2013.0, "completions/mean_length": 1624.09375, "completions/mean_terminated_length": 732.6060180664062, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.9838695911922847, "frac_reward_zero_std": 0.75, "grad_norm": 26.341631261954728, "learning_rate": 1.0076653295123537e-07, "loss": -0.0015, "num_tokens": 2718267191.0, "reward": 0.0546875, "reward_std": 0.0893455445766449, "rewards/accuracy_reward/mean": 0.0546875, "rewards/accuracy_reward/std": 0.2275916188955307, "step": 2882 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.74609375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2043.0, "completions/mean_length": 1690.67578125, "completions/mean_terminated_length": 640.6923217773438, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.9842109755056755, "frac_reward_zero_std": 0.75, "grad_norm": 62.39520358592612, "learning_rate": 1.0073557358722834e-07, "loss": 0.0154, "num_tokens": 2719215185.0, "reward": 0.037109375, "reward_std": 0.08412160724401474, "rewards/accuracy_reward/mean": 0.038306452333927155, "rewards/accuracy_reward/std": 0.19212883710861206, "step": 2883 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.802734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1942.0, "completions/mean_length": 1779.228515625, "completions/mean_terminated_length": 685.5148315429688, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.9845523598190663, "frac_reward_zero_std": 0.84375, "grad_norm": 37.69118220102959, "learning_rate": 1.0070525187190301e-07, "loss": -0.0062, "num_tokens": 2720212166.0, "reward": 0.02734375, "reward_std": 0.05311024561524391, "rewards/accuracy_reward/mean": 0.02734375, "rewards/accuracy_reward/std": 0.16324250400066376, "step": 2884 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.71875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2018.0, "completions/mean_length": 1639.041015625, "completions/mean_terminated_length": 593.9236450195312, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.9848937441324571, "frac_reward_zero_std": 0.78125, "grad_norm": 41.51337757613398, "learning_rate": 1.0067556784829557e-07, "loss": -0.009, "num_tokens": 2721134459.0, "reward": 0.05859375, "reward_std": 0.08450747281312943, "rewards/accuracy_reward/mean": 0.05859375, "rewards/accuracy_reward/std": 0.23509246110916138, "step": 2885 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.685546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1600.986328125, "completions/mean_terminated_length": 626.4409790039062, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9852351284458479, "frac_reward_zero_std": 0.75, "grad_norm": 23.35618071605464, "learning_rate": 1.0064652155853695e-07, "loss": 0.0222, "num_tokens": 2722034676.0, "reward": 0.0703125, "reward_std": 0.094524085521698, "rewards/accuracy_reward/mean": 0.0703125, "rewards/accuracy_reward/std": 0.25592297315597534, "step": 2886 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.654296875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2040.0, "completions/mean_length": 1576.390625, "completions/mean_terminated_length": 683.796630859375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9855765127592387, "frac_reward_zero_std": 0.84375, "grad_norm": 106.26054271929235, "learning_rate": 1.0061811304385314e-07, "loss": 0.0058, "num_tokens": 2722916876.0, "reward": 0.03515625, "reward_std": 0.05385598540306091, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 2887 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.68359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1917.0, "completions/mean_length": 1580.5, "completions/mean_terminated_length": 570.4691162109375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9859178970726296, "frac_reward_zero_std": 0.71875, "grad_norm": 55.7578031730014, "learning_rate": 1.0059034234456476e-07, "loss": 0.0212, "num_tokens": 2723803324.0, "reward": 0.07421875, "reward_std": 0.11943789571523666, "rewards/accuracy_reward/mean": 0.07421875, "rewards/accuracy_reward/std": 0.2623828947544098, "step": 2888 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.720703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1948.0, "completions/mean_length": 1640.810546875, "completions/mean_terminated_length": 590.0908813476562, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.9862592813860203, "frac_reward_zero_std": 0.78125, "grad_norm": 11.58924987969007, "learning_rate": 1.0056320950008728e-07, "loss": 0.0008, "num_tokens": 2724726091.0, "reward": 0.03125, "reward_std": 0.0737205371260643, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 2889 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.740234375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1958.0, "completions/mean_length": 1662.82421875, "completions/mean_terminated_length": 565.2180786132812, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9866006656994111, "frac_reward_zero_std": 0.8125, "grad_norm": 16.871163938001057, "learning_rate": 1.0053671454893084e-07, "loss": 0.002, "num_tokens": 2725653249.0, "reward": 0.05078125, "reward_std": 0.07113292813301086, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 2890 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2045.0, "completions/mean_length": 1564.34765625, "completions/mean_terminated_length": 608.2907104492188, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9869420500128019, "frac_reward_zero_std": 0.71875, "grad_norm": 19.01770263631309, "learning_rate": 1.0051085752870009e-07, "loss": -0.0184, "num_tokens": 2726532147.0, "reward": 0.07421875, "reward_std": 0.11180429905653, "rewards/accuracy_reward/mean": 0.07421875, "rewards/accuracy_reward/std": 0.2623828947544098, "step": 2891 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.689453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2023.0, "completions/mean_length": 1610.35546875, "completions/mean_terminated_length": 638.7295532226562, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9872834343261927, "frac_reward_zero_std": 0.90625, "grad_norm": 12.516958522735711, "learning_rate": 1.0048563847609443e-07, "loss": 0.0173, "num_tokens": 2727435081.0, "reward": 0.033203125, "reward_std": 0.0234375, "rewards/accuracy_reward/mean": 0.03427419438958168, "rewards/accuracy_reward/std": 0.18211629986763, "step": 2892 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2016.0, "completions/mean_length": 1569.640625, "completions/mean_terminated_length": 590.1428833007812, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9876248186395835, "frac_reward_zero_std": 0.75, "grad_norm": 86.34302535217995, "learning_rate": 1.0046105742690761e-07, "loss": 0.0158, "num_tokens": 2728321905.0, "reward": 0.052734375, "reward_std": 0.10155140608549118, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 2893 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.783203125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2008.0, "completions/mean_length": 1740.587890625, "completions/mean_terminated_length": 630.0270385742188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9879662029529743, "frac_reward_zero_std": 0.78125, "grad_norm": 141.25029384303645, "learning_rate": 1.0043711441602797e-07, "loss": 0.0061, "num_tokens": 2729288862.0, "reward": 0.033203125, "reward_std": 0.07493096590042114, "rewards/accuracy_reward/mean": 0.033203125, "rewards/accuracy_reward/std": 0.17934183776378632, "step": 2894 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.767578125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1972.0, "completions/mean_length": 1730.060546875, "completions/mean_terminated_length": 680.058837890625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9883075872663651, "frac_reward_zero_std": 0.6875, "grad_norm": 20.967274329792332, "learning_rate": 1.0041380947743828e-07, "loss": 0.0124, "num_tokens": 2730253725.0, "reward": 0.09765625, "reward_std": 0.12328367680311203, "rewards/accuracy_reward/mean": 0.09765625, "rewards/accuracy_reward/std": 0.29713961482048035, "step": 2895 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.68359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1949.0, "completions/mean_length": 1595.552734375, "completions/mean_terminated_length": 618.043212890625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.988648971579756, "frac_reward_zero_std": 0.75, "grad_norm": 53.658209639368096, "learning_rate": 1.0039114264421555e-07, "loss": 0.0023, "num_tokens": 2731140536.0, "reward": 0.05078125, "reward_std": 0.09962804615497589, "rewards/accuracy_reward/mean": 0.05078125, "rewards/accuracy_reward/std": 0.21976542472839355, "step": 2896 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.708984375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1847.0, "completions/mean_length": 1639.779296875, "completions/mean_terminated_length": 645.2550659179688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9889903558931467, "frac_reward_zero_std": 0.6875, "grad_norm": 16.241572705096488, "learning_rate": 1.0036911394853133e-07, "loss": 0.0471, "num_tokens": 2732053495.0, "reward": 0.064453125, "reward_std": 0.1035923957824707, "rewards/accuracy_reward/mean": 0.064453125, "rewards/accuracy_reward/std": 0.24579854309558868, "step": 2897 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1909.0, "completions/mean_length": 1637.783203125, "completions/mean_terminated_length": 666.2171020507812, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9893317402065375, "frac_reward_zero_std": 0.90625, "grad_norm": 0.753228457685946, "learning_rate": 1.0034772342165116e-07, "loss": 0.0088, "num_tokens": 2732980232.0, "reward": 0.03125, "reward_std": 0.031083684414625168, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 2898 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.62109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1917.0, "completions/mean_length": 1517.029296875, "completions/mean_terminated_length": 646.6752319335938, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.9896731245199283, "frac_reward_zero_std": 0.78125, "grad_norm": 85.30452265228864, "learning_rate": 1.003269710939351e-07, "loss": 0.0151, "num_tokens": 2733827559.0, "reward": 0.04296875, "reward_std": 0.07972928881645203, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 2899 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.693359375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1858.0, "completions/mean_length": 1600.865234375, "completions/mean_terminated_length": 589.8280639648438, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9900145088333191, "frac_reward_zero_std": 0.78125, "grad_norm": 38.20320033151139, "learning_rate": 1.0030685699483725e-07, "loss": 0.0014, "num_tokens": 2734721266.0, "reward": 0.060546875, "reward_std": 0.07438573241233826, "rewards/accuracy_reward/mean": 0.060546875, "rewards/accuracy_reward/std": 0.2387305200099945, "step": 2900 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.732421875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2006.0, "completions/mean_length": 1678.359375, "completions/mean_terminated_length": 666.5693359375, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.9903558931467099, "frac_reward_zero_std": 0.875, "grad_norm": 2.2930164611398363, "learning_rate": 1.0028738115290592e-07, "loss": -0.0012, "num_tokens": 2735661338.0, "reward": 0.015625, "reward_std": 0.04081955552101135, "rewards/accuracy_reward/mean": 0.015625, "rewards/accuracy_reward/std": 0.12414088100194931, "step": 2901 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1956.0, "completions/mean_length": 1598.896484375, "completions/mean_terminated_length": 610.8687744140625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.9906972774601007, "frac_reward_zero_std": 0.84375, "grad_norm": 5.315972544352134, "learning_rate": 1.0026854359578345e-07, "loss": 0.0074, "num_tokens": 2736558021.0, "reward": 0.017578125, "reward_std": 0.04957009106874466, "rewards/accuracy_reward/mean": 0.017578125, "rewards/accuracy_reward/std": 0.13154059648513794, "step": 2902 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1953.0, "completions/mean_length": 1669.76953125, "completions/mean_terminated_length": 624.4705810546875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9910386617734915, "frac_reward_zero_std": 0.84375, "grad_norm": 2.951849838514443, "learning_rate": 1.0025034435020633e-07, "loss": 0.0141, "num_tokens": 2737497231.0, "reward": 0.01953125, "reward_std": 0.04907120764255524, "rewards/accuracy_reward/mean": 0.01953125, "rewards/accuracy_reward/std": 0.1385180652141571, "step": 2903 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.685546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1834.0, "completions/mean_length": 1567.134765625, "completions/mean_terminated_length": 518.788818359375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.9913800460868823, "frac_reward_zero_std": 0.78125, "grad_norm": 47.18804606024116, "learning_rate": 1.0023278344200509e-07, "loss": 0.0076, "num_tokens": 2738378884.0, "reward": 0.056640625, "reward_std": 0.0870293527841568, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 2904 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.677734375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1998.0, "completions/mean_length": 1598.876953125, "completions/mean_terminated_length": 654.3575439453125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9917214304002732, "frac_reward_zero_std": 0.8125, "grad_norm": 38.53162246213055, "learning_rate": 1.0021586089610422e-07, "loss": 0.0064, "num_tokens": 2739266485.0, "reward": 0.021484375, "reward_std": 0.05688370764255524, "rewards/accuracy_reward/mean": 0.021484375, "rewards/accuracy_reward/std": 0.14513419568538666, "step": 2905 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.685546875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1748.0, "completions/mean_length": 1598.623046875, "completions/mean_terminated_length": 618.9254760742188, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9920628147136639, "frac_reward_zero_std": 0.65625, "grad_norm": 10.001976236915034, "learning_rate": 1.0019957673652214e-07, "loss": 0.03, "num_tokens": 2740164692.0, "reward": 0.126953125, "reward_std": 0.15081490576267242, "rewards/accuracy_reward/mean": 0.126953125, "rewards/accuracy_reward/std": 0.33324605226516724, "step": 2906 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.720703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1956.0, "completions/mean_length": 1660.46484375, "completions/mean_terminated_length": 660.4615478515625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9924041990270547, "frac_reward_zero_std": 0.6875, "grad_norm": 20.51033762081488, "learning_rate": 1.0018393098637125e-07, "loss": 0.0185, "num_tokens": 2741093090.0, "reward": 0.09765625, "reward_std": 0.13130201399326324, "rewards/accuracy_reward/mean": 0.09765625, "rewards/accuracy_reward/std": 0.29713961482048035, "step": 2907 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.69140625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2000.0, "completions/mean_length": 1602.169921875, "completions/mean_terminated_length": 603.2848510742188, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9927455833404455, "frac_reward_zero_std": 0.625, "grad_norm": 49.746729670366946, "learning_rate": 1.0016892366785786e-07, "loss": 0.0041, "num_tokens": 2741987561.0, "reward": 0.076171875, "reward_std": 0.12961195409297943, "rewards/accuracy_reward/mean": 0.076171875, "rewards/accuracy_reward/std": 0.26553234457969666, "step": 2908 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.720703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2048.0, "completions/mean_length": 1666.03125, "completions/mean_terminated_length": 680.3916015625, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "epoch": 0.9930869676538363, "frac_reward_zero_std": 0.65625, "grad_norm": 25.34545420345999, "learning_rate": 1.0015455480228208e-07, "loss": 0.006, "num_tokens": 2742921081.0, "reward": 0.052734375, "reward_std": 0.11993582546710968, "rewards/accuracy_reward/mean": 0.052734375, "rewards/accuracy_reward/std": 0.22372129559516907, "step": 2909 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.720703125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2037.0, "completions/mean_length": 1654.248046875, "completions/mean_terminated_length": 638.2028198242188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9934283519672271, "frac_reward_zero_std": 0.71875, "grad_norm": 124.40137879401742, "learning_rate": 1.0014082441003791e-07, "loss": 0.0184, "num_tokens": 2743854008.0, "reward": 0.0703125, "reward_std": 0.11641617119312286, "rewards/accuracy_reward/mean": 0.0703125, "rewards/accuracy_reward/std": 0.25592297315597534, "step": 2910 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.755859375, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.0, "completions/mean_length": 1712.30859375, "completions/mean_terminated_length": 673.416015625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.9937697362806179, "frac_reward_zero_std": 0.78125, "grad_norm": 136.7744002463834, "learning_rate": 1.001277325106131e-07, "loss": 0.001, "num_tokens": 2744813702.0, "reward": 0.064453125, "reward_std": 0.08825473487377167, "rewards/accuracy_reward/mean": 0.06653226166963577, "rewards/accuracy_reward/std": 0.2494617998600006, "step": 2911 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.7109375, "completions/max_length": 2048.0, "completions/max_terminated_length": 1841.0, "completions/mean_length": 1633.568359375, "completions/mean_terminated_length": 614.29052734375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9941111205940087, "frac_reward_zero_std": 0.90625, "grad_norm": 0.7808250238000148, "learning_rate": 1.0011527912258924e-07, "loss": 0.015, "num_tokens": 2745725305.0, "reward": 0.0234375, "reward_std": 0.03449726849794388, "rewards/accuracy_reward/mean": 0.0234375, "rewards/accuracy_reward/std": 0.15143637359142303, "step": 2912 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.73828125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1916.0, "completions/mean_length": 1687.01171875, "completions/mean_terminated_length": 668.7014770507812, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9944525049073996, "frac_reward_zero_std": 0.6875, "grad_norm": 62.25508818880001, "learning_rate": 1.0010346426364161e-07, "loss": 0.0128, "num_tokens": 2746659631.0, "reward": 0.0625, "reward_std": 0.11795367300510406, "rewards/accuracy_reward/mean": 0.06451612710952759, "rewards/accuracy_reward/std": 0.2459181249141693, "step": 2913 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.62890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2047.0, "completions/mean_length": 1511.3125, "completions/mean_terminated_length": 601.7684326171875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.9947938892207903, "frac_reward_zero_std": 0.78125, "grad_norm": 34.70397896268805, "learning_rate": 1.0009228795053926e-07, "loss": 0.021, "num_tokens": 2747510831.0, "reward": 0.056640625, "reward_std": 0.08384227007627487, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 2914 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2045.0, "completions/mean_length": 1587.34765625, "completions/mean_terminated_length": 644.107177734375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9951352735341811, "frac_reward_zero_std": 0.8125, "grad_norm": 50.38357555931986, "learning_rate": 1.0008175019914494e-07, "loss": -0.0072, "num_tokens": 2748395553.0, "reward": 0.04296875, "reward_std": 0.07477159798145294, "rewards/accuracy_reward/mean": 0.04296875, "rewards/accuracy_reward/std": 0.2029850035905838, "step": 2915 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.705078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1871.0, "completions/mean_length": 1644.50390625, "completions/mean_terminated_length": 679.8543090820312, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9954766578475719, "frac_reward_zero_std": 0.78125, "grad_norm": 46.49427255684848, "learning_rate": 1.0007185102441505e-07, "loss": 0.0045, "num_tokens": 2749323075.0, "reward": 0.037109375, "reward_std": 0.07894542813301086, "rewards/accuracy_reward/mean": 0.037109375, "rewards/accuracy_reward/std": 0.18921469151973724, "step": 2916 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.6875, "completions/max_length": 2048.0, "completions/max_terminated_length": 1983.0, "completions/mean_length": 1605.68359375, "completions/mean_terminated_length": 632.5875244140625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.9958180421609627, "frac_reward_zero_std": 0.8125, "grad_norm": 43.39794387579375, "learning_rate": 1.0006259044039964e-07, "loss": 0.0075, "num_tokens": 2750221601.0, "reward": 0.041015625, "reward_std": 0.0673912987112999, "rewards/accuracy_reward/mean": 0.041015625, "rewards/accuracy_reward/std": 0.19852031767368317, "step": 2917 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.697265625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1967.0, "completions/mean_length": 1627.2109375, "completions/mean_terminated_length": 658.0386962890625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9961594264743535, "frac_reward_zero_std": 0.78125, "grad_norm": 33.95512716338215, "learning_rate": 1.000539684602424e-07, "loss": -0.005, "num_tokens": 2751126301.0, "reward": 0.025390625, "reward_std": 0.06755761802196503, "rewards/accuracy_reward/mean": 0.025390625, "rewards/accuracy_reward/std": 0.15746226906776428, "step": 2918 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.640625, "completions/max_length": 2048.0, "completions/max_terminated_length": 2046.0, "completions/mean_length": 1521.2890625, "completions/mean_terminated_length": 582.3695678710938, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9965008107877443, "frac_reward_zero_std": 0.625, "grad_norm": 19.961791525407527, "learning_rate": 1.0004598509618068e-07, "loss": 0.0082, "num_tokens": 2751981857.0, "reward": 0.107421875, "reward_std": 0.14617595076560974, "rewards/accuracy_reward/mean": 0.107421875, "rewards/accuracy_reward/std": 0.30995169281959534, "step": 2919 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.689453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1979.0, "completions/mean_length": 1600.60546875, "completions/mean_terminated_length": 607.654052734375, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.9968421951011351, "frac_reward_zero_std": 0.6875, "grad_norm": 114.35284707076231, "learning_rate": 1.0003864035954539e-07, "loss": 0.0071, "num_tokens": 2752891511.0, "reward": 0.08203125, "reward_std": 0.11687077581882477, "rewards/accuracy_reward/mean": 0.08203125, "rewards/accuracy_reward/std": 0.2746807038784027, "step": 2920 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.69921875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2013.0, "completions/mean_length": 1619.150390625, "completions/mean_terminated_length": 622.2142944335938, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.997183579414526, "frac_reward_zero_std": 0.65625, "grad_norm": 15.177119997846791, "learning_rate": 1.0003193426076107e-07, "loss": 0.0146, "num_tokens": 2753796020.0, "reward": 0.03515625, "reward_std": 0.09853477776050568, "rewards/accuracy_reward/mean": 0.03515625, "rewards/accuracy_reward/std": 0.1843547374010086, "step": 2921 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.666015625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1937.0, "completions/mean_length": 1562.55859375, "completions/mean_terminated_length": 594.5146484375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9975249637279167, "frac_reward_zero_std": 0.71875, "grad_norm": 47.36340697254016, "learning_rate": 1.0002586680934577e-07, "loss": 0.0121, "num_tokens": 2754673874.0, "reward": 0.056640625, "reward_std": 0.1076337993144989, "rewards/accuracy_reward/mean": 0.056640625, "rewards/accuracy_reward/std": 0.23138070106506348, "step": 2922 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.63671875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2028.0, "completions/mean_length": 1516.759765625, "completions/mean_terminated_length": 585.6613159179688, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "epoch": 0.9978663480413075, "frac_reward_zero_std": 0.84375, "grad_norm": 8.409006334902793, "learning_rate": 1.0002043801391112e-07, "loss": 0.0062, "num_tokens": 2755520679.0, "reward": 0.03125, "reward_std": 0.05641176179051399, "rewards/accuracy_reward/mean": 0.03125, "rewards/accuracy_reward/std": 0.17416280508041382, "step": 2923 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.716796875, "completions/max_length": 2048.0, "completions/max_terminated_length": 2031.0, "completions/mean_length": 1676.609375, "completions/mean_terminated_length": 736.6068725585938, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9982077323546983, "frac_reward_zero_std": 0.6875, "grad_norm": 31.8641102083544, "learning_rate": 1.0001564788216237e-07, "loss": 0.0199, "num_tokens": 2756449871.0, "reward": 0.07421875, "reward_std": 0.12037497013807297, "rewards/accuracy_reward/mean": 0.07421875, "rewards/accuracy_reward/std": 0.2623828947544098, "step": 2924 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.712890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1977.0, "completions/mean_length": 1663.767578125, "completions/mean_terminated_length": 709.7210693359375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "epoch": 0.9985491166680891, "frac_reward_zero_std": 0.5625, "grad_norm": 73.80738272060073, "learning_rate": 1.0001149642089817e-07, "loss": -0.0022, "num_tokens": 2757381752.0, "reward": 0.07421875, "reward_std": 0.1546139121055603, "rewards/accuracy_reward/mean": 0.07421875, "rewards/accuracy_reward/std": 0.2623828947544098, "step": 2925 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.705078125, "completions/max_length": 2048.0, "completions/max_terminated_length": 2004.0, "completions/mean_length": 1630.35546875, "completions/mean_terminated_length": 631.8807983398438, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.9988905009814799, "frac_reward_zero_std": 0.75, "grad_norm": 68.42868504750174, "learning_rate": 1.0000798363601074e-07, "loss": 0.0277, "num_tokens": 2758291102.0, "reward": 0.046875, "reward_std": 0.09039659798145294, "rewards/accuracy_reward/mean": 0.046875, "rewards/accuracy_reward/std": 0.21157780289649963, "step": 2926 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.728515625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1882.0, "completions/mean_length": 1681.779296875, "completions/mean_terminated_length": 699.0431518554688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9992318852948707, "frac_reward_zero_std": 0.9375, "grad_norm": 8.45734651403176, "learning_rate": 1.0000510953248593e-07, "loss": -0.0037, "num_tokens": 2759226093.0, "reward": 0.009765625, "reward_std": 0.023271184414625168, "rewards/accuracy_reward/mean": 0.009765625, "rewards/accuracy_reward/std": 0.09843364357948303, "step": 2927 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.689453125, "completions/max_length": 2048.0, "completions/max_terminated_length": 1920.0, "completions/mean_length": 1603.537109375, "completions/mean_terminated_length": 616.7735595703125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.9995732696082615, "frac_reward_zero_std": 0.78125, "grad_norm": 15.172487258378336, "learning_rate": 1.0000287411440292e-07, "loss": -0.0058, "num_tokens": 2760117328.0, "reward": 0.0546875, "reward_std": 0.07730703055858612, "rewards/accuracy_reward/mean": 0.05645161122083664, "rewards/accuracy_reward/std": 0.23102475702762604, "step": 2928 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.712890625, "completions/max_length": 2048.0, "completions/max_terminated_length": 1999.0, "completions/mean_length": 1657.44921875, "completions/mean_terminated_length": 687.7142944335938, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "epoch": 0.9999146539216524, "frac_reward_zero_std": 0.78125, "grad_norm": 38.57795061124567, "learning_rate": 1.0000127738493448e-07, "loss": -0.0021, "num_tokens": 2761048070.0, "reward": 0.0703125, "reward_std": 0.08351518213748932, "rewards/accuracy_reward/mean": 0.0703125, "rewards/accuracy_reward/std": 0.25592297315597534, "step": 2929 }, { "epoch": 0.9999146539216524, "step": 2929, "total_flos": 0.0, "train_loss": 0.012068198455147701, "train_runtime": 103778.9718, "train_samples_per_second": 0.903, "train_steps_per_second": 0.028 } ], "logging_steps": 1, "max_steps": 2930, "num_input_tokens_seen": 2761048070, "num_train_epochs": 1, "save_steps": 147, "stateful_callbacks": { "TrainerControl": { "args": { "should_epoch_stop": false, "should_evaluate": false, "should_log": false, "should_save": true, "should_training_stop": false }, "attributes": {} } }, "total_flos": 0.0, "train_batch_size": 16, "trial_name": null, "trial_params": null }