MAI-2024-0049

The Heuristic Token Search HTS Attack represents a sophisticated method for circumventing safety protocols in text-to-image T2I models, enabling the generation of Not Safe For Work NSFW content. This attack involves systematically substituting tokens within a malicious prompt with semantically...

Backdoor Attack on Vision Language Models with Stealthy Semantic Manipulation

Vision Language Models VLMs have shown remarkable performance, but are also vulnerable to backdoor attacks whereby the adversary can manipulate the model's outputs through hidden triggers. Prior attacks primarily rely on single-modality triggers, leaving the crucial cross-modal fusion nature of...