Beware the Rationalization Trap! When Language Model Explainability Diverges from our Mental Models of Language

Open in new window