---
dataset_info:
  features:
  - name: chosen
    list:
    - name: content
      dtype: string
    - name: role
      dtype: string
  - name: rejected
    list:
    - name: content
      dtype: string
    - name: role
      dtype: string
  - name: chosen_rating
    dtype: float64
  - name: rejected_rating
    dtype: float64
  - name: metadata
    dtype: string
  - name: chosen_rewards
    sequence: float32
  - name: rejected_rewards
    sequence: float32
  splits:
  - name: train
    num_bytes: 7926924
    num_examples: 2418
  download_size: 3080003
  dataset_size: 7926924
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---