mlsanityData to 7 Oct 2026

Learning to Detect Unknown Jailbreak Attacks in Large Vision-Language Models

arXiv 2510.15430 ↗First posted 17 Oct 2025cs.CV cs.AIcited by 1

On the map: Multimodal LLMs / vision-language models → MLLM jailbreaks and adversarial robustness

also in 1 more
Abstract (arXiv)

Despite extensive alignment efforts, Large Vision-Language Models (LVLMs) remain vulnerable to jailbreak attacks, posing serious safety risks. To address this, existing detection methods either learn attack-specific parameters, which hinders generalization to unseen attacks, or rely on heuristically sound principles, which limit accuracy and efficiency. To overcome these limitations, we propose Learning to Detect (LoD), a general framework that accurately detects unknown jailbreak attacks by shifting the focus from attack-specific learning to task-specific learning. This framework includes a Multi-modal Safety Concept Activation Vector module for safety-oriented representation learning and a Safety Pattern Auto-Encoder module for unsupervised attack classification. Extensive experiments show that our method achieves consistently higher detection AUROC on diverse unknown attacks while improving efficiency. The code is available at https://anonymous.4open.science/r/Learning-to-Detect-51CB.

Key facts. Cited by 1 paper in the map (data through 7 Oct 2026). Placed in MLLM jailbreaks and adversarial robustness, one of 16 directions in Multimodal LLMs / vision-language models. Builds on 21 of its 36 references that are in the map. First posted on arXiv 17 Oct 2025.

Builds on

21 of 36 references.

Closest ancestors · 10

Same direction, newest first

  1. 2025HiddenDetect: Detecting Jailbreak Attacks against Large Vision-Language Models via Monitoring Hidden States20 Feb
  2. 2025A Survey of Safety on Large Vision-Language Models: Attacks, Defenses and Evaluations14 Feb
  3. 2024JailbreakZoo: Survey, Landscapes, and Horizons in Jailbreaking Large Language and Vision-Language Models26 Jun
  4. 2024Unbridled Icarus: A Survey of the Potential Perils of Image Inputs in Multimodal Large Language Model Security8 Apr
  5. 2024JailBreakV: A Benchmark for Assessing the Robustness of MultiModal Large Language Models against Jailbreak Attacks3 Apr
  6. 2024Eyes Closed, Safety On: Protecting Multimodal LLMs via Image-to-Text Transformation14 Mar
  7. 2024MLLM-Protector: Ensuring MLLM's Safety without Hurting Performance5 Jan
  8. 2023FigStep: Jailbreaking Large Vision-Language Models via Typographic Visual Prompts9 Nov
  9. 2023Jailbreak in pieces: Compositional Adversarial Attacks on Multi-Modal Language Models26 Jul
  10. 2023Visual Adversarial Examples Jailbreak Aligned Large Language Models22 Jun

Other references · 11

  1. 2023MM-Vet: Evaluating Large Multimodal Models for Integrated Capabilitiescited by 1,221
  2. 2023Detecting Language Model Attacks with Perplexitycited by 411
  3. 2023ToxicChat: Unveiling Hidden Challenges of Toxicity Detection in Real-World User-AI Conversationcited by 224
  4. 2022Why Should Adversarial Perturbations be Imperceptible? Rethink the Research Paradigm in Adversarial NLPcited by 99
  5. 2024GradSafe: Detecting Jailbreak Prompts for LLMs via Safety-Critical Gradient Analysiscited by 87
  6. 2024Uncovering Safety Risks of Large Language Models through Concept Activation Vectorcited by 63
  7. 2024MM-Vet v2: A Challenging Benchmark to Evaluate Large Multimodal Models for Integrated Capabilitiescited by 57
  8. 2024Jailbreak Large Vision-Language Models Through Multi-Modal Linkagecited by 53
  9. 2025Playing the Fool: Jailbreaking LLMs and Multimodal LLMs with Out-of-Distribution Strategycited by 37
  10. 2025MOAT: Evaluating LMMs for Capability Integration and Instruction Groundingcited by 2
  11. 2025A Comparative Study of Rule-Based and Data-Driven Approaches in Industrial Monitoringcited by 2

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2026: 1.

1 paper.

  1. 2026Safe responses matter: Output-aware safety guardrail mitigate over-refusal in MLLMscited by 2

Similar, not cited

Close in wording, no citation either way

  1. 2025Learning to Detect Unseen Jailbreak Attacks in Large Vision-Language Models
  2. 2025JailDAM: Jailbreak Detection with Adaptive Memory for Vision-Language Model
  3. 2025Rethinking Jailbreak Detection of Large Vision Language Models with Representational Contrastive Scoring
  4. 2024ImgTrojan: Jailbreaking Vision-Language Models with ONE Image
  5. 2025PRISM: Programmatic Reasoning with Image Sequence Manipulation for LVLM Jailbreaking
  6. 2025SafeLLM: Unlearning Harmful Outputs from Large Language Models against Jailbreak Attacks
  7. 2025JBShield: Defending Large Language Models from Jailbreak Attacks through Activated Concept Analysis and Manipulation
  8. 2025Defense-to-Attack: Bypassing Weak Defenses Enables Stronger Jailbreaks in Vision-Language Models