---
dataset_info:
  features:
  - name: prompt
    dtype: string
  - name: chosen
    dtype: string
  - name: rejected
    dtype: string
  - name: chosen_score
    dtype: float64
  - name: rejected_score
    dtype: float64
  splits:
  - name: train
    num_bytes: 8048555
    num_examples: 4603
  download_size: 4810679
  dataset_size: 8048555
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---