---
dataset_info:
  features:
  - name: conversations
    sequence:
    - name: value
      dtype: string
    - name: source
      dtype: string
  - name: chosen
    struct:
    - name: value
      dtype: string
    - name: source
      dtype: string
  - name: rejected
    struct:
    - name: value
      dtype: string
    - name: source
      dtype: string
  - name: task_id
    dtype: int32
  - name: step_id
    dtype: int32
  - name: source_name
    dtype: string
  - name: image_url
    dtype: string
  splits:
  - name: train
    num_bytes: 316853156
    num_examples: 30921
  download_size: 32134792
  dataset_size: 316853156
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---