metadata
dataset_info:
features:
- name: instruction
dtype: string
- name: response_a
dtype: string
- name: response_b
dtype: string
- name: label
dtype: int64
splits:
- name: train
num_bytes: 538358
num_examples: 500
download_size: 117249
dataset_size: 538358
configs:
- config_name: default
data_files:
- split: train
path: data/train-*