metadata
dataset_info:
features:
- name: prompt
dtype: string
- name: classes
sequence: string
- name: answer_index
dtype: int32
- name: round
dtype: int32
splits:
- name: train
num_bytes: 13766772
num_examples: 20897
download_size: 2293543
dataset_size: 13766772
configs:
- config_name: default
data_files:
- split: train
path: data/train-*