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: prompt
dtype: string
splits:
- name: train
num_bytes: 36074746
num_examples: 26277
- name: test
num_bytes: 3938751
num_examples: 2824
download_size: 22663180
dataset_size: 40013497
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*