metadata
dataset_info:
features:
- name: chosen
list:
- name: role
dtype: string
- name: content
dtype: string
- name: rejected
list:
- name: role
dtype: string
- name: content
dtype: string
- name: source
dtype: string
splits:
- name: train
num_bytes: 100960065.14743547
num_examples: 51200
- name: test
num_bytes: 9859381.362054246
num_examples: 5000
download_size: 61894430
dataset_size: 110819446.50948972
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*