Pairing a decoy image with an encoded jailbreak prompt can sharply amplify the safety effect of an image-aware defense (ECSO), at the cost of more benign refusals unless gated by a detector.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CR 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Decoy Images Amplify Caption-Mediated Defenses Against Encoded Jailbreaks
Pairing a decoy image with an encoded jailbreak prompt can sharply amplify the safety effect of an image-aware defense (ECSO), at the cost of more benign refusals unless gated by a detector.