metadata
dataset_info:
features:
- name: prompt
dtype: string
- name: chosen
dtype: string
- name: rejected
dtype: string
- name: __index_level_0__
dtype: int64
splits:
- name: train
num_bytes: 20845030
num_examples: 13286
- name: test
num_bytes: 2293753
num_examples: 1477
download_size: 10822837
dataset_size: 23138783
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*