---
dataset_info:
  features:
  - name: prompt
    dtype: string
  - name: answers
    sequence: string
  - name: label
    sequence: int64
  - name: step_scores
    sequence: float64
  splits:
  - name: train
    num_bytes: 767766990
    num_examples: 1319
  download_size: 145318010
  dataset_size: 767766990
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---