---
dataset_info:
  features:
  - name: answers
    sequence: string
  - name: gt
    dtype: string
  - name: first_reward
    dtype: bool
  - name: prompt_messages
    list:
    - name: content
      dtype: string
    - name: role
      dtype: string
  splits:
  - name: train
    num_bytes: 409228915
    num_examples: 44968
  download_size: 137864349
  dataset_size: 409228915
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---