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: annotator_1
dtype: int64
- name: annotator_2
dtype: int64
- name: annotator_3
dtype: int64
- name: winner
dtype: string
- name: turn
dtype: int64
splits:
- name: validation
num_bytes: 1051003
num_examples: 999
download_size: 145870
dataset_size: 1051003
configs:
- config_name: default
data_files:
- split: validation
path: data/validation-*