1 paper
Lingyun Zhang, Yu Xie, Ping Chen
The nature of personalized text-to-image models poses a unique safety challenge that generic context-blind methods are ill-equipped to handle. Such global filters create a dilemma:…