---
dataset_info:
  features:
  - name: policy_prompt
    dtype: string
  - name: winner_response
    dtype: string
  - name: loser_response
    dtype: string
  splits:
  - name: train
    num_bytes: 83200
    num_examples: 1000
  - name: validation
    num_bytes: 5280
    num_examples: 100
  download_size: 4755
  dataset_size: 88480
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: validation
    path: data/validation-*
---