metadata
dataset_info:
features:
- name: id
dtype: string
- name: source_id
dtype: string
- name: page_id
dtype: string
- name: qa_type
dtype: string
- name: answer_letter
dtype: string
- name: messages
list:
- name: role
dtype: string
- name: content
dtype: string
splits:
- name: train
num_bytes: 3428307
num_examples: 3198
- name: test
num_bytes: 181170
num_examples: 169
download_size: 3357079
dataset_size: 3609477
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*