Adversarial Training for Multimodal Large Language Models against Jailbreak Attacks

AAML
Re-assign community