metadata
dataset_info:
features:
- name: rejected
dtype: string
- name: chosen
dtype: string
- name: prompt
dtype: string
splits:
- name: train
num_bytes: 2740378
num_examples: 2109
download_size: 1009085
dataset_size: 2740378
configs:
- config_name: default
data_files:
- split: train
path: data/train-*