Revealing the Parametric Knowledge of Language Models: A Unified Framework for Attribution Methods

Open in new window