metadata
dataset_info:
features:
- name: chosen
dtype: string
- name: rejected
dtype: string
- name: group_id
dtype: int64
- name: prompt
dtype: string
splits:
- name: train
num_bytes: 33782752
num_examples: 7944
download_size: 4220169
dataset_size: 33782752
configs:
- config_name: default
data_files:
- split: train
path: data/train-*