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: 29885404
num_examples: 21968
- name: test
num_bytes: 3273949
num_examples: 2362
download_size: 18800601
dataset_size: 33159353
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*