commonsense reasoning

We introduce the novel Wino-X benchmark to investigate whether translation models can perform coreference resolution that requires commonsense knowledge and whether multilingual language models are capable of commonsense reasoning across multiple languages. Our findings indicate that models are prone to biases and often fail to identify disambiguating information.

We investingate the ability of neural and classification models to reason about (im)moral behavior grounded in concrete, structured, social situations.