Zhifang Zhang

TokenSwap: Backdoor Attack on the Compositional Understanding of Large Vision-Language Models

Improving Generalizability and Undetectability for Targeted Adversarial Attacks on Multimodal Pre-trained Models

Solving Puzzles? Jailbreaking Multimodal Large Language Models!

A Closer Look at Backdoor Attacks on CLIP

Defending Multimodal Backdoored Models by Repulsive Visual Prompt Tuning

A Closer Look at Backdoor Attacks on CLIP