metadata
dataset_info:
features:
- name: qid
dtype: string
- name: chosen
list:
- name: content
dtype: string
- name: role
dtype: string
- name: rejected
list:
- name: content
dtype: string
- name: role
dtype: string
splits:
- name: train
num_bytes: 14256438583.53102
num_examples: 3798835
- name: test
num_bytes: 1584049565.4689815
num_examples: 422093
download_size: 8864292488
dataset_size: 15840488149
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*