metadata
dataset_info:
features:
- name: chosen
dtype: string
- name: rejected
dtype: string
- name: prompt
dtype: string
- name: __index_level_0__
dtype: int64
splits:
- name: train
num_bytes: 6250299
num_examples: 10000
download_size: 1981558
dataset_size: 6250299
configs:
- config_name: default
data_files:
- split: train
path: data/train-*