---
dataset_info:
  features:
  - name: prompt
    list:
    - name: content
      dtype: string
    - name: role
      dtype: string
  - name: chosen
    list:
    - name: content
      dtype: string
    - name: role
      dtype: string
  - name: rejected
    list:
    - name: content
      dtype: string
    - name: role
      dtype: string
  - name: images
    struct:
    - name: bytes
      dtype: binary
    - name: path
      dtype: 'null'
  splits:
  - name: train
    num_bytes: 21329611
    num_examples: 30
  download_size: 21295061
  dataset_size: 21329611
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---