AI Paper Atlas

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

Towards Robust Multimodal Large Language Models Against Jailbreak Attacks

Ziyi Yin, Yuanpu Cao, Han Liu, Ting Wang, Jinghui Chen, Fenglong Ma

CVPR 2026 · Adversarial robustness and security › Attacks on text and graphs · approach: LLM jailbreak and injection defense · 4 citations (OpenAlex)

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
adversarial tuning with SafeMLLMMLLMs (6 models) · ImgJP (AdvBench)ASR: 51.33 -> 5.17Improved
adversarial tuning with SafeMLLMMLLMs (6 models) · GCG (AdvBench)ASR: 33.83 -> 0Improved
adversarial tuning with SafeMLLM instead of CATMLLMs (6 models) · FigStep (SafeBench-Tiny)ASR: 19.33 -> 1.00Improved
SafeMLLM instead of VLGuard safety tuningLLaVA-1.5-7B · ImgJP (AdvBench)ASR: 88.00 -> 6.00Improved
SafeMLLM instead of VLGuardLLaVA-1.5-7B · FigStepASR: 2.00 -> 0.00No clear effect
add contrastive loss to attack and defense (ablation)SafeMLLM · ImgJP (AdvBench)ASR: 23.00 -> 0.00Improved
add utility loss on benign pairs (ablation)SafeMLLM · MM-Vetscore: 21.6 -> 37.8Improved
add utility loss on benign pairs (ablation)SafeMLLM · MM-Vet (MiniGPT-4)score: 7.2 -> 22.8Improved
add Ph0 and Pt0 token embedding perturbations (vs removing) (ablation)SafeMLLM · ImgJP, LLaVA-1.5ASRNo clear effect
token-embedding perturbation instead of adversarial image (ablation)SafeMLLM · LLaVA-1.5-7Bruntime per iteration: 84.42 -> 20.73Improved
token-embedding perturbation instead of latent adversarial training (ablation)SafeMLLM · LLaVA-1.5-13Bruntime per iteration: 192.39 -> 38.70Improved

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