Research map / Large language models / Language, safety and interpretability
Efficient LLM-Jailbreaking via Multimodal-LLM Jailbreak
Haoxuan Ji, Zheng Lin, Zhenxing Niu, Xinbo Gao, Gang Hua
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 changed | Where | Measured | Effect |
|---|---|---|---|
| jailbreak LLM via MLLM jailbreak embedding instead of GCG token search | LLaMA2-Chat-7B · AdvBench-M test | avg ASR (%): 74.7 -> 84.7 | Improved |
| use MLLM-jailbreak approach instead of AdvAttack | InternLM2.5-7B · AdvBench-M test | avg ASR (%): 58.0 -> 84.6 | Improved |
| use MLLM approach instead of GCG for runtime | LLaMA2-Chat-7B · per-class runtime | hours: 11.2h -> 0.37h | Improved |
| use image-text semantic matching for initial image instead of random (ablation) | MLLM jailbreak · AdvBench-M Dtrain (MiniGPT4) | ASR (%): 73.32 -> 83.70 | Improved |
| use image-text semantic matching instead of ranking-based init (ablation) | MLLM jailbreak · AdvBench-M Dtrain (MiniGPT4) | ASR (%): 75.18 -> 83.70 | Improved |
| optimize image via visual module instead of directly optimizing soft embeddings (ablation) | embedding-based jailbreak · token conversion | cosine similarity: 0.17 -> 0.63 | Improved |
| compare ours with AdvAttack on validation split | InternLM2.5-7B · HarmBench VAL | ASR (%): 68.7 -> 59.0 | Hurt |
| transfer InternVL2.0-derived suffixes to proprietary models | black-box targets · AdvBench-M | avg ASR (%): 77.1 (Mistral) vs 20.1 (DouBao) | Mixed |
Closest papers by meaning
- An Interpretable N-gram Perplexity Threat Model for Large Language Model Jailbreaks (ICML 2025 · similarity 0.91)
- Efficient LLM Jailbreak via Adaptive Dense-to-sparse Constrained Optimization (NeurIPS 2024 · similarity 0.91)
- MMJ-Bench: A Comprehensive Study on Jailbreak Attacks and Defenses for Vision Language Models (AAAI 2025 · similarity 0.90)
- Weak-to-Strong Jailbreaking on Large Language Models (ICML 2025 · similarity 0.90)
- Improved Few-Shot Jailbreaking Can Circumvent Aligned Language Models and Their Defenses (NeurIPS 2024 · similarity 0.90)
- JailbreakBench: An Open Robustness Benchmark for Jailbreaking Large Language Models (NeurIPS 2024 · similarity 0.90)
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.
