metadata
dataset_info:
features:
- name: prompt
dtype: string
- name: chosen
dtype: string
- name: rejected
dtype: string
- name: preference
dtype: string
- name: reason
dtype: string
splits:
- name: train
num_bytes: 11790
num_examples: 10
download_size: 12249
dataset_size: 11790
configs:
- config_name: default
data_files:
- split: train
path: data/train-*