---
dataset_info:
  features:
  - name: model_a
    dtype: string
  - name: model_b
    dtype: string
  - name: prompt
    dtype: string
  - name: response_a
    dtype: string
  - name: conversation_a
    list:
    - name: content
      dtype: string
    - name: role
      dtype: string
  - name: response_b
    dtype: string
  - name: conversation_b
    list:
    - name: content
      dtype: string
    - name: role
      dtype: string
  - name: winner
    dtype: string
  splits:
  - name: train
    num_bytes: 383079019
    num_examples: 57477
  download_size: 206868425
  dataset_size: 383079019
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---