metadata
dataset_info:
features:
- name: prompt
dtype: string
- name: score
dtype: string
- name: chosen
dtype: string
- name: rejected
dtype: string
- name: __index_level_0__
dtype: int64
splits:
- name: train
num_bytes: 35855
num_examples: 42
download_size: 30782
dataset_size: 35855
configs:
- config_name: default
data_files:
- split: train
path: data/train-*