metadata
dataset_info:
features:
- name: prompt
dtype: int64
- name: chosen
dtype: int64
- name: rejected
dtype: int64
splits:
- name: train
num_bytes: 24
num_examples: 1
download_size: 1850
dataset_size: 24
configs:
- config_name: default
data_files:
- split: train
path: data/train-*