---
dataset_info:
  features:
  - name: prompt
    list:
    - name: content
      dtype: string
    - name: role
      dtype: string
  - name: a1
    list:
    - name: content
      dtype: string
    - name: role
      dtype: string
  - name: a2
    list:
    - name: content
      dtype: string
    - name: role
      dtype: string
  - name: rank
    dtype: int64
  splits:
  - name: train
    num_bytes: 107000127
    num_examples: 43835
  download_size: 64503950
  dataset_size: 107000127
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---