---
dataset_info:
  features:
  - name: message_response_cut_id
    dtype: string
  - name: completion_input_ids
    sequence: int32
  - name: rollout_gen_params
    struct:
    - name: max_tokens
      dtype: int32
    - name: temperature
      dtype: float32
    - name: top_p
      dtype: float32
  - name: batch_id
    dtype: string
  splits:
  - name: train
    num_bytes: 2402760900
    num_examples: 301696
  download_size: 674426058
  dataset_size: 2402760900
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---