mlsanityData to 14 Oct 2026

LLaVAShield: Safeguarding Multimodal Multi-Turn Dialogues in Vision-Language Models

arXiv 2509.25896 ↗First posted 30 Sep 2025cs.CVcited by 1

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

Abstract (arXiv)

As Vision-Language Models (VLMs) move into interactive, multi-turn use, safety concerns intensify for multimodal multi-turn dialogue, which is characterized by concealment of malicious intent, contextual risk accumulation, and cross-modal joint risk. These characteristics limit the effectiveness of content moderation approaches designed for single-turn or single-modality settings. To address these limitations, we first construct the Multimodal Multi-turn Dialogue Safety (MMDS) dataset, comprising 4,484 annotated dialogues and a comprehensive risk taxonomy with 8 primary and 60 subdimensions. As part of MMDS construction, we introduce Multimodal Multi-turn Red Teaming (MMRT), an automated framework for generating unsafe multimodal multi-turn dialogues. We further propose LLaVAShield, which audits the safety of both user inputs and assistant responses under specified policy dimensions in multimodal multi-turn dialogues. Extensive experiments show that LLaVAShield significantly outperforms state-of-the-art VLMs and existing content moderation tools while demonstrating strong generalization and flexible policy adaptation. Additionally, we analyze vulnerabilities of mainstream VLMs to harmful inputs and evaluate the contribution of key components, advancing understanding of safety mechanisms in multimodal multi-turn dialogues.

Key facts. Cited by 1 paper 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 60 of its 82 references that are in the map. First posted on arXiv 30 Sep 2025.

Builds on

60 of 82 references.

Closest ancestors · 16

Same direction, newest first

  1. 2025Unveiling Trust in Multimodal Large Language Models: Evaluation, Analysis, and Mitigation21 Aug
  2. 2025Visual Contextual Attack: Jailbreaking MLLMs with Image-Driven Context Injection3 Jul
  3. 2025SafeCoT: Improving VLM Safety with Minimal Reasoning10 Jun
  4. 2025VisCRA: A Visual Chain Reasoning Attack for Jailbreaking Multimodal Large Language Models26 May
  5. 2025Survey of Adversarial Robustness in Multimodal Large Language Models18 Mar
  6. 2025Can't See the Forest for the Trees: Benchmarking Multimodal Safety Awareness for Multimodal LLMs16 Feb
  7. 2024RapGuard: Safeguarding Multimodal Large Language Models via Rationale-aware Defensive Prompting25 Dec
  8. 2024Llama Guard 3 Vision: Safeguarding Human-AI Image Understanding Conversations15 Nov
  9. 2024IDEATOR: Jailbreaking and Benchmarking Large Vision-Language Models Using Themselves29 Oct
  10. 2024Unraveling and Mitigating Safety Alignment Degradation of Vision-Language Models11 Oct
  11. 2024MLLMGuard: A Multi-dimensional Safety Evaluation Suite for Multimodal Large Language Models11 Jun
  12. 2024MultiTrust: A Comprehensive Benchmark Towards Trustworthy Multimodal Large Language Models11 Jun
4 more
  1. 2024Safety Alignment for Vision Language Models22 May
  2. 2024JailBreakV: A Benchmark for Assessing the Robustness of MultiModal Large Language Models against Jailbreak Attacks3 Apr
  3. 2023On the Robustness of Large Multimodal Models Against Image Adversarial Attacks6 Dec
  4. 2023MM-SafetyBench: A Benchmark for Safety Evaluation of Multimodal Large Language Models29 Nov

Other references · 44

  1. 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,744
  2. 2025Qwen3 Technical Reportcited by 8,820
  3. 2025Qwen2.5-VL Technical Reportcited by 5,587
  4. 2024GPT-4o System Cardcited by 4,588
  5. 2024Scaling Rectified Flow Transformers for High-Resolution Image Synthesiscited by 4,181
  6. 2025Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilitiescited by 3,796
  7. 2023Universal and Transferable Adversarial Attacks on Aligned Language Modelscited by 3,104
  8. 2024LLaVA-OneVision: Easy Visual Task Transfercited by 2,719
  9. 2023InternVL: Scaling up Vision Foundation Models and Aligning for Generic Visual-Linguistic Taskscited by 2,565
  10. 2025InternVL3: Exploring Advanced Training and Test-Time Recipes for Open-Source Multimodal Modelscited by 1,834
  11. 2025InternVL3.5: Advancing Open-Source Multimodal Models in Versatility, Reasoning, and Efficiencycited by 1,510
  12. 2023Jailbreaking Black Box Large Language Models in Twenty Queriescited by 1,420
32 more
  1. 2023Llama Guard: LLM-based Input-Output Safeguard for Human-AI Conversationscited by 1,207
  2. 2024WildChat: 1M ChatGPT Interaction Logs in the Wildcited by 629
  3. 2023GPTFUZZER: Red Teaming Large Language Models with Auto-Generated Jailbreak Promptscited by 560
  4. 2024WildGuard: Open One-Stop Moderation Tools for Safety Risks, Jailbreaks, and Refusals of LLMscited by 496
  5. 2024Great, Now Write an Article About That: The Crescendo Multi-Turn LLM Jailbreak Attackcited by 390
  6. 2022A Holistic Approach to Undesired Content Detection in the Real Worldcited by 356
  7. 2020Learning from the Worst: Dynamically Generated Datasets to Improve Online Hate Detectioncited by 221
  8. 2022A New Generation of Perspective API: Efficient Multilingual Character-level Transformerscited by 185
  9. 2022Talking About Large Language Modelscited by 175
  10. 2024Ovis: Structural Embedding Alignment for Multimodal Large Language Modelcited by 169
  11. 2024Safety Fine-Tuning at (Almost) No Cost: A Baseline for Vision Large Language Modelscited by 167
  12. 2025GuardReasoner: Towards Reasoning-based LLM Safeguardscited by 100
  13. 2024MMDU: A Multi-Turn Multi-Image Dialog Understanding Benchmark and Instruction-Tuning Dataset for LVLMscited by 90
  14. 2024CogView3: Finer and Faster Text-to-Image Generation via Relay Diffusioncited by 72
  15. 2024SafeBench: A Safety Evaluation Framework for Multimodal Large Language Modelscited by 69
  16. 2024Speak Out of Turn: Safety Vulnerability of Large Language Models in Multi-turn Dialoguecited by 66
  17. 2025Reasoning-Augmented Conversation for Multi-Turn Jailbreak Attacks on Large Language Modelscited by 64
  18. 2024Jailbreak Large Vision-Language Models Through Multi-Modal Linkagecited by 62
  19. 2024Multi-Turn Context Jailbreak Attack on Large Language Models From First Principlescited by 56
  20. 2024UnsafeBench: Benchmarking Image Safety Classifiers on Real-World and AI-Generated Imagescited by 53
  21. 2024LlavaGuard: An Open VLM-based Framework for Safeguarding Vision Datasets and Modelscited by 52
  22. 2024CoSafe: Evaluating Large Language Model Safety in Multi-Turn Dialogue Coreferencecited by 51
  23. 2025GuardReasoner-VL: Safeguarding VLMs via Reinforced Reasoningcited by 42
  24. 2025ShieldGemma 2: Robust and Tractable Image Content Moderationcited by 28
  25. 2025SafeDialBench: A Fine-Grained Safety Evaluation Benchmark for Large Language Models in Multi-Turn Dialogues with Diverse Jailbreak Attackscited by 23
  26. 2025BingoGuard: LLM Content Moderation Tools with Risk Levelscited by 21
  27. 2025Breaking the Ceiling: Exploring the Potential of Jailbreak Attacks through Expanding Strategy Spacecited by 17
  28. 2025Multi-Turn Jailbreaks Are Simpler Than They Seemcited by 17
  29. 2025MMCR: Advancing Visual Language Model in Multimodal Multi-Turn Contextual Reasoningcited by 16
  30. 2025ShieldVLM: Safeguarding the Multimodal Implicit Toxicity via Deliberative Reasoning with LVLMscited by 7
  31. 2025SafeTy Reasoning Elicitation Alignment for Multi-Turn Dialoguescited by 7
  32. 2025Many-Turn Jailbreakingcited by 2

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 1.

1 paper.

  1. 2026Visual Memory Injection Attacks for Multi-Turn Conversationscited by 2

Similar, not cited

Close in wording, no citation either way

  1. 2025SafeMT: Multi-turn Safety for Multimodal Language Models
  2. 2026MTMCS-Bench: Evaluating Contextual Safety of Multimodal Large Language Models in Multi-Turn Dialogues
  3. 2026Towards Multi-modal Multi-turn Safety: From Agentic Interaction to Strategic Alignment
  4. 2024Multimodal Situational Safety
  5. 2025MSTS: A Multimodal Safety Test Suite for Vision-Language Models
  6. 2026Lingua-SafetyBench: A Benchmark for Safety Evaluation of Multilingual Vision-Language Models
  7. 2025MSR-Align: Policy-Grounded Multimodal Alignment for Safety-Aware Reasoning in Vision-Language Models
  8. 2025RedDiffuser: Auditing Multimodal Safety Failures in Vision-Language Models via Reinforced Diffusion