JRQi/geomlama

Data from the paper GeoMLAMA: Geo-Diverse Commonsense Probing on Multilingual Pre-Trained Language Models (The 125 raw instances are combined with the 500 augmented instances (without "Corresponding Questions")) GitHub