Explicit image-tool interaction in VLMs cuts multimodal jailbreak ASR by ~30% on average; the effect is attributed to a safety-relevant shift in hidden representations rather than image semantics or text traces.
Jailbreaks on vision language model via multimodal reasoning
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
When Think-with-Image Meets Safety: What Determines Multimodal Jailbreak Robustness?
Explicit image-tool interaction in VLMs cuts multimodal jailbreak ASR by ~30% on average; the effect is attributed to a safety-relevant shift in hidden representations rather than image semantics or text traces.