metadata
dataset_info:
features:
- name: prompt
dtype: string
- name: response_a
dtype: string
- name: response_b
dtype: string
- name: more_reasonable
dtype: string
splits:
- name: train
num_bytes: 913012
num_examples: 3426
download_size: 209984
dataset_size: 913012
configs:
- config_name: default
data_files:
- split: train
path: data/train-*