metadata
dataset_info:
features:
- name: chosen
list:
- name: content
dtype: string
- name: role
dtype: string
- name: rejected
list:
- name: content
dtype: string
- name: role
dtype: string
- name: domain
dtype: 'null'
splits:
- name: train
num_bytes: 17407060
num_examples: 4046
- name: test
num_bytes: 1566669
num_examples: 352
download_size: 9283199
dataset_size: 18973729
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*