metadata
dataset_info:
features:
- name: prompt
dtype: string
- name: response_a
dtype: string
- name: response_b
dtype: string
- name: more_reasonable
dtype: string
- name: __index_level_0__
dtype: int64
splits:
- name: train
num_bytes: 23557
num_examples: 100
download_size: 15532
dataset_size: 23557
configs:
- config_name: default
data_files:
- split: train
path: data/train-*