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: 17223
num_examples: 125
- name: fa
num_bytes: 24061
num_examples: 125
- name: hi
num_bytes: 34719
num_examples: 125
- name: sw
num_bytes: 17593
num_examples: 125
- name: zh
num_bytes: 15926
num_examples: 125
- name: el
num_bytes: 37639
num_examples: 150
download_size: 45285
dataset_size: 147161
data from the paper GeoMLAMA: Geo-Diverse Commonsense Probing on Multilingual Pre-Trained Language Models (along with some new data and modifications for cleaning)