---
dataset_info:
  features:
  - name: prompt
    dtype: string
  - name: ground_truth
    dtype: string
  - name: base_response
    dtype: string
  splits:
  - name: train
    num_bytes: 89126866
    num_examples: 5395
  download_size: 30434763
  dataset_size: 89126866
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---