āGeographic and Geopolitical Biases of Language Modelsā, 2022-12-20 (; backlinks)ā :
Pretrained language models (PLMs) often fail to fairly represent target users from certain world regions because of the under-representation of those regions in training datasets. With recent PLMs trained on enormous data sources, quantifying their potential biases is difficult, due to their black-box nature and the sheer scale of the data sources.
In this work, we devise an approach to study the geographic bias (and knowledge) present in PLMs [GPT-2 & BLOOM], proposing a Geographic-Representation Probing Framework adopting a self-conditioning method coupled with entity-country mappings.
Our findings suggest PLMsā representations map surprisingly well to the physical world in terms of country-to-country associations, but this knowledge is unequally shared across languages. Last, we explain how large PLMs despite exhibiting notions of geographical proximity, over-amplify geopolitical favoritism at inference time.