mlsanityData to 14 Oct 2026

UniGuard: Towards Universal Safety Guardrails for Jailbreak Attacks on Multimodal Large Language Models

arXiv 2411.01703 ↗First posted 3 Nov 2024cs.CL cs.AI cs.LGcited by 19

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

also in 1 more
Abstract (arXiv)

Multimodal large language models (MLLMs) have revolutionized vision-language understanding but remain vulnerable to multimodal jailbreak attacks, where adversarial inputs are meticulously crafted to elicit harmful or inappropriate responses. We propose UniGuard, a novel multimodal safety guardrail that jointly considers the unimodal and cross-modal harmful signals. UniGuard trains a multimodal guardrail to minimize the likelihood of generating harmful responses in a toxic corpus. The guardrail can be seamlessly applied to any input prompt during inference with minimal computational costs. Extensive experiments demonstrate the generalizability of UniGuard across multiple modalities, attack strategies, and multiple state-of-the-art MLLMs, including LLaVA, Gemini Pro, GPT-4o, MiniGPT-4, and InstructBLIP. Notably, this robust defense mechanism maintains the models' overall vision-language understanding capabilities.

Key facts. Cited by 19 papers in the map (data through 14 Oct 2026). Placed in MLLM jailbreaks and adversarial robustness, one of 16 directions in Multimodal LLMs / vision-language models. Builds on 39 of its 53 references that are in the map. First posted on arXiv 3 Nov 2024.

Builds on

39 of 53 references.

Closest ancestors · 3

Same direction, newest first

  1. 2024Eyes Closed, Safety On: Protecting Multimodal LLMs via Image-to-Text Transformation14 Mar
  2. 2024MLLM-Protector: Ensuring MLLM's Safety without Hurting Performance5 Jan
  3. 2023Jailbreak in pieces: Compositional Adversarial Attacks on Multi-Modal Language Models26 Jul

Other references · 36

  1. 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,744
  2. 2019PyTorch: An Imperative Style, High-Performance Deep Learning Librarycited by 20,305
  3. 2014Microsoft COCO: Common Objects in Contextcited by 17,608
  4. 2023Llama 2: Open Foundation and Fine-Tuned Chat Modelscited by 13,028
  5. 2022Chain-of-Thought Prompting Elicits Reasoning in Large Language Modelscited by 12,459
  6. 2023Visual Instruction Tuningcited by 7,869
  7. 2017Towards Deep Learning Models Resistant to Adversarial Attackscited by 6,415
  8. 2023Improved Baselines with Visual Instruction Tuningcited by 4,418
  9. 2021The Power of Scale for Parameter-Efficient Prompt Tuningcited by 3,795
  10. 2024Gemini 1.5: Unlocking multimodal understanding across millions of tokens of contextcited by 3,224
  11. 2023Universal and Transferable Adversarial Attacks on Aligned Language Modelscited by 3,104
  12. 2023InstructBLIP: Towards General-purpose Vision-Language Models with Instruction Tuningcited by 2,876
24 more
  1. 2023MiniGPT-4: Enhancing Vision-Language Understanding with Advanced Large Language Modelscited by 2,711
  2. 2023Jailbreaking Black Box Large Language Models in Twenty Queriescited by 1,420
  3. 2020RealToxicityPrompts: Evaluating Neural Toxic Degeneration in Language Modelscited by 1,255
  4. 2023MM-Vet: Evaluating Large Multimodal Models for Integrated Capabilitiescited by 1,239
  5. 2022A-OKVQA: A Benchmark for Visual Question Answering using World Knowledgecited by 710
  6. 2022EVA: Exploring the Limits of Masked Visual Representation Learning at Scalecited by 691
  7. 2023OpenFlamingo: An Open-Source Framework for Training Large Autoregressive Vision-Language Modelscited by 565
  8. 2022Diffusion Models for Adversarial Purificationcited by 476
  9. 2023SmoothLLM: Defending Large Language Models Against Jailbreaking Attackscited by 469
  10. 2023Image Hijacks: Adversarial Images can Control Generative Models at Runtimecited by 172
  11. 2024Safety Fine-Tuning at (Almost) No Cost: A Baseline for Vision Large Language Modelscited by 167
  12. 2019Sparse and Imperceivable Adversarial Attackscited by 95
  13. 2023Better to Ask in English: Cross-Lingual Evaluation of Large Language Models for Healthcare Queriescited by 82
  14. 2024When Search Engine Services meet Large Language Models: Visions and Challengescited by 59
  15. 2023MUFFIN: Curating Multi-Faceted Instructions for Improving Instruction-Followingcited by 30
  16. 2023Tackling Data Bias in MUSIC-AVQA: Crafting a Balanced Dataset for Unbiased Question-Answeringcited by 29
  17. 2024MM-Soc: Benchmarking Multimodal Large Language Models in Social Media Platformscited by 26
  18. 2024Deconstructing The Ethics of Large Language Models from Long-standing Issues to New-emerging Dilemmas: A Surveycited by 23
  19. 2023MuseChat: A Conversational Music Recommendation System for Videoscited by 21
  20. 2023Online Advertisements with LLMs: Opportunities and Challengescited by 17
  21. 2023G2uardFL: Safeguarding Federated Learning Against Backdoor Attacks through Attributed Client Graph Clusteringcited by 9
  22. 2023Adversarial Robustness of Prompt-based Few-Shot Learning for Natural Language Understandingcited by 8
  23. 2023Foundation Model-oriented Robustness: Robust Image Model Evaluation with Pretrained Modelscited by 6
  24. 2023Pointing out Human Answer Mistakes in a Goal-Oriented Visual Dialoguecited by 1

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2024: 1; 2025: 12; 2026: 6.

19 papers.

Sort:

By area:

  1. 2025Large Language Model Agent: A Survey on Methodology, Applications and Challengescited by 205
  2. 2025A Survey of Safety on Large Vision-Language Models: Attacks, Defenses and Evaluationscited by 67
  3. 2025VPI-Bench: Visual Prompt Injection Attacks for Computer-Use Agentscited by 45
  4. 2025GuardReasoner-VL: Safeguarding VLMs via Reinforced Reasoningcited by 42
  5. 2024Jailbreak Attacks and Defenses against Multimodal Generative Models: A Surveycited by 32
  6. 2025Understanding and Rectifying Safety Perception Distortion in VLMscited by 17
  7. 2025No Free Lunch with Guardrailscited by 11
  8. 2025OmniSafeBench-MM: A Unified Benchmark and Toolbox for Multimodal Jailbreak Attack-Defense Evaluationcited by 11
  9. 2025Beyond Magic Words: Sharpness-Aware Prompt Evolving for Robust Large Language Models with TAREcited by 9
  10. 2025DAVSP: Safety Alignment for Large Vision-Language Models via Deep Aligned Visual Safety Promptcited by 8
  11. 2025EigenShield: Causal Subspace Filtering via Random Matrix Theory for Adversarially Robust Vision-Language Modelscited by 2
  12. 2026Risk Awareness Injection: Calibrating Vision-Language Models for Safety without Compromising Utilitycited by 2
7 more
  1. 2025ScreenLLM: Stateful Screen Schema for Efficient Action Understanding and Predictioncited by 1
  2. 2025Beyond Text: Multimodal Jailbreaking of Vision-Language and Audio Models through Perceptually Simple Transformationscited by 1
  3. 2026Feature-Space Smoothing: Certified Robustness of Deep Representationscited by 0
  4. 2026SaFeR-ToolKit: Structured Reasoning via Virtual Tool Calling for Multimodal Safetycited by 0
  5. 2026Securing Multimodal AI through Internal Information Decompositioncited by 0
  6. 2026Evolving Safety Landscape of Multi-modal Large Language Models: A Survey of Emerging Threats and Safeguardscited by 0
  7. 2026SafeRI: Recognition and Intervention for Token-Level Safety Intervention in Large Vision Language Modelscited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2025Towards Effective MLLM Jailbreaking Through Balanced On-Topicness and OOD-Intensity
  2. 2025HiddenDetect: Detecting Jailbreak Attacks against Large Vision-Language Models via Monitoring Hidden States
  3. 2025Align is not Enough: Multimodal Universal Jailbreak Attack against Multimodal Large Language Models
  4. 2026Jailbreaks on Vision Language Model via Multimodal Reasoning
  5. 2025Con Instruction: Universal Jailbreaking of Multimodal Large Language Models via Non-Textual Modalities
  6. 2026Text-Anchored Semantic Perturbations for Transferable Jailbreak Attacks on Multimodal Large Language Models
  7. 2025PolyJailbreak: Cross-Modal Jailbreaking Attacks on Black-Box Multimodal LLMs
  8. 2026Toward Universal and Transferable Jailbreak Attacks on Vision-Language Models