---
dataset_info:
  features:
  - name: prompt
    dtype: string
  - name: chosen
    dtype: string
  - name: rejected
    dtype: string
  - name: model
    dtype: string
  - name: chosen_score
    dtype: float64
  - name: rejected_score
    dtype: float64
  - name: margin
    dtype: float64
  splits:
  - name: train
    num_bytes: 221003934
    num_examples: 58368
  download_size: 123932549
  dataset_size: 221003934
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---