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: reason
dtype: string
splits:
- name: train
num_bytes: 101159141.35300794
num_examples: 5021
- name: test
num_bytes: 5339010.646992055
num_examples: 265
download_size: 23857693
dataset_size: 106498152
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*