eval_results.json

{
    "epoch": 0.9981298423724285,
    "eval_logits/chosen": -1.535585641860962,
    "eval_logits/rejected": -1.557785153388977,
    "eval_logps/chosen": -1.0460667610168457,
    "eval_logps/rejected": -1.119951605796814,
    "eval_loss": 1.6345860958099365,
    "eval_rewards/accuracies": 0.5685483813285828,
    "eval_rewards/chosen": -2.6151669025421143,
    "eval_rewards/margins": 0.18471203744411469,
    "eval_rewards/rejected": -2.7998790740966797,
    "eval_runtime": 42.9398,
    "eval_samples": 1961,
    "eval_samples_per_second": 45.669,
    "eval_steps_per_second": 1.444
}