---
dataset_info:
  features:
  - name: idx
    dtype: int64
  - name: prompt
    dtype: string
  - name: answers
    sequence: string
  - name: gt
    dtype: string
  - name: rewards
    sequence: float64
  splits:
  - name: train
    num_bytes: 2928500854
    num_examples: 20000
  download_size: 1084919296
  dataset_size: 2928500854
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---