AI Paper Atlas

Research map / Adversarial robustness and security / Attacks on text and graphs

SafeLogo: Turning Your Logos into Jailbreak Shields via Micro-Regional Adversarial Training

Zhiyi Duan, Xiaoyue Zhang, Tianxing Man

CVPR 2026 · Adversarial robustness and security › Attacks on text and graphs · approach: Vision-language model robustness

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add SafeLogo on top of a safety instruction (ablation)LLaVA-1.5-13B · MM-SafetyBench SD+TYPO GCGASR (%): 12.8% -> 3.1%Improved
use SafeLogo alone without safety instruction (ablation)LLaVA-1.5-13B · MM-SafetyBench SD+TYPO GCGASR (%): 17.1% -> 18.7%No clear effect
increase defense weight from 50% to 98% (ablation)SafeLogo · MM-SafetyBench / MM-VetASR (%) / MM-Vet: ASR 31% -> 6%; MM-Vet 39.9 -> 35.3Mixed
raise defense weight to 99% (ablation)SafeLogo · MM-SafetyBenchASR (%): 6% -> 8%Hurt
increase logo coverage from 1% to 9% (ablation)SafeLogo · MM-SafetyBenchASR (%): 29% -> 5%Improved
increase coverage from 9% to 16% (ablation)SafeLogo · MM-SafetyBench / MM-VetASR (%) / MM-Vet: ASR 5% -> 8%; MM-Vet 38.5 -> 37.0Hurt
perturbation range 64/255 versus 96/255 (ablation)SafeLogo · MM-SafetyBench / MM-VetASR (%) / MM-Vet: ASR 8% -> 15%; MM-Vet 39.1 -> 36.5 at 96/255Hurt
perturbation range 128/255 versus 64/255 (ablation)SafeLogo · MM-SafetyBench / MM-VetASR (%) / MM-Vet: ASR 8% -> 7%; MM-Vet 39.1 -> 37.2Mixed
full-intensity perturbation 255/255 (ablation)SafeLogo · MM-SafetyBenchASR (%): 8% -> 22%Hurt
apply LLaVA-trained SafeLogo to Gemini-2.5-Flash-ImageSafeLogo · VLGuard (Gemini-2.5-Flash-Image)ASR (%): 2.9/4.8/1.0% -> 1.0/0.0/0.0%Improved
apply SafeLogo to Qwen3-VL-8BQwen3-VL-8B-Instruct · MM-Vet totalscore: 65.8 -> 53.3Hurt
apply SafeLogo to LLaVA-1.5-13BLLaVA-1.5-13B · MME total (OOD)score: 1820 -> 1553Hurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy