metadata
dataset_info:
features:
- name: question
dtype: string
- name: response
dtype: string
- name: correct_alternative
dtype: string
- name: prediction
dtype: string
splits:
- name: train
num_bytes: 181150
num_examples: 84
download_size: 130828
dataset_size: 181150
configs:
- config_name: default
data_files:
- split: train
path: data/train-*