Universal adversarial attacks cause output perturbation 90 times more often than precise target injection in VLMs, with only 2 verbatim successes out of 6615 tests.
Universal adversarial attack on aligned multimodal llms
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
citation-role summary
background 1
citation-polarity summary
verdicts
CONDITIONAL 2roles
background 1polarities
unclear 1representative citing papers
Generative purification with consensus aggregation reduces adversarial illusion attack success rates to near zero on ImageBind while improving alignment on both clean and attacked inputs.
citing papers explorer
-
VisInject: Disruption != Injection -- A Dual-Dimension Evaluation of Universal Adversarial Attacks on Vision-Language Models
Universal adversarial attacks cause output perturbation 90 times more often than precise target injection in VLMs, with only 2 verbatim successes out of 6615 tests.