How Smart is BERT? Evaluating the Language Model’s Commonsense Knowledge

Researchers dive deep into the large language model to discover how it encodes the structured commonsense knowledge it leverages on downstream commonsense tasks.