1 paper · 1 filter
Keyu Li, Jin Gao, Jialing Zhang +1
Concept unlearning is increasingly used to limit the reproduction of protected or unsafe visual concepts in text-to-image models. Existing evaluations, however, mostly study target…