metadata
dataset_info:
features:
- name: question
dtype: string
- name: answer
dtype: string
- name: candidate_answers
dtype: string
- name: context
dtype: string
- name: country
dtype: string
splits:
- name: en
num_bytes: 20705
num_examples: 150
- name: fa
num_bytes: 29418
num_examples: 150
- name: hi
num_bytes: 41903
num_examples: 150
- name: sw
num_bytes: 21231
num_examples: 150
- name: zh
num_bytes: 19155
num_examples: 150
- name: el
num_bytes: 38057
num_examples: 150
download_size: 45566
dataset_size: 170469
data from the paper GeoMLAMA: Geo-Diverse Commonsense Probing on Multilingual Pre-Trained Language Models (along with some new data and modifications for cleaning)