ReFACT: Updating Text-to-Image Models by Editing the Text Encoder.

Dana Arad, Hadas Orgad, Yonatan Belinkov

Browse the full NAACL paper archive.