JaiLIP: Jailbreaking Vision-Language Models via Loss Guided Image Perturbation.

Md. Jueal Mia, M. Hadi Amini

Browse the full ICMLA paper archive.