---
dataset_info:
  features:
  - name: data_source
    dtype: string
  - name: prompt
    list:
    - name: content
      dtype: string
    - name: role
      dtype: string
  - name: reward_model
    struct:
    - name: ground_truth
      dtype: string
    - name: style
      dtype: string
  - name: question
    dtype: string
  - name: solution
    dtype: string
  - name: answer
    dtype: string
  splits:
  - name: train
    num_bytes: 26406244
    num_examples: 42398
  - name: valid
    num_bytes: 915137
    num_examples: 1513
  download_size: 14136451
  dataset_size: 27321381
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: valid
    path: data/valid-*
---