metadata
dataset_info:
features:
- name: conversation_a
list:
- name: content
dtype: string
- name: role
dtype: string
- name: conversation_b
list:
- name: content
dtype: string
- name: role
dtype: string
- name: winner
dtype: string
- name: turn
dtype: int64
splits:
- name: validation
num_bytes: 233427
num_examples: 80
download_size: 138209
dataset_size: 233427
configs:
- config_name: default
data_files:
- split: validation
path: data/validation-*