metadata
dataset_info:
features:
- name: doc_id
dtype: string
- name: type
dtype: string
- name: question
dtype: string
- name: answer
dtype: string
- name: evidence_pages
dtype: string
- name: evidence_sources
dtype: string
- name: answer_format
dtype: string
splits:
- name: train
num_bytes: 222517
num_examples: 1062
download_size: 81055
dataset_size: 222517
configs:
- config_name: default
data_files:
- split: train
path: data/train-*