# Think in Safety: Unveiling and Mitigating Safety Alignment Collapse in Multimodal Large Reasoning Model

arXiv: https://arxiv.org/abs/2505.06538 · HTML: https://mlsanity.com/paper/2505.06538/

## Key facts

- Cited by 19 papers in the map (data through 14 Oct 2026).
- Placed in Reasoning for agents, planning and code, one of 17 directions in LLM reasoning.
- Builds on 33 of its 49 references that are in the map.
- First posted on arXiv 10 May 2025.

## Placed in

- LLM reasoning / [Reasoning for agents, planning and code](https://mlsanity.com/lr/reasoning-for-agents-planning-and-code/)
- LLM alignment / RLHF / preference optimisation / [Safety alignment and jailbreak attacks](https://mlsanity.com/alignment/safety-alignment-and-jailbreak-attacks/)
- LLM safety and security / [Safety alignment, refusal and guardrails](https://mlsanity.com/safety/safety-alignment-refusal-and-guardrails/)

## Builds on (most cited first, 30 of 33)

- [DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Models](https://mlsanity.com/paper/2402.03300/) (2024)
- [DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning](https://mlsanity.com/paper/2501.12948/) (2025)
- [Qwen2.5-VL Technical Report](https://mlsanity.com/paper/2502.13923/) (2025)
- [GPT-4o System Card](https://mlsanity.com/paper/2410.21276/) (2024)
- [OpenAI o1 System Card](https://mlsanity.com/paper/2412.16720/) (2024)
- [DeepSeek-V3 Technical Report](https://mlsanity.com/paper/2412.19437/) (2024)
- [Vision-R1: Incentivizing Reasoning Capability in Multimodal Large Language Models](https://mlsanity.com/paper/2503.06749/) (2025)
- [MLE-bench: Evaluating Machine Learning Agents on Machine Learning Engineering](https://mlsanity.com/paper/2410.07095/) (2024)
- [LLaVA-CoT: Let Vision Language Models Reason Step-by-Step](https://mlsanity.com/paper/2411.10440/) (2024)
- [Kimi-VL Technical Report](https://mlsanity.com/paper/2504.07491/) (2025)
- [R1-Onevision: Advancing Generalized Multimodal Reasoning through Cross-Modal Formalization](https://mlsanity.com/paper/2503.10615/) (2025)
- [Deliberative Alignment: Reasoning Enables Safer Language Models](https://mlsanity.com/paper/2412.16339/) (2024)
- [MM-Eureka: Exploring the Frontiers of Multimodal Reasoning with Rule-based Reinforcement Learning](https://mlsanity.com/paper/2503.07365/) (2025)
- [LMM-R1: Empowering 3B LMMs with Strong Reasoning Abilities Through Two-Stage Rule-Based RL](https://mlsanity.com/paper/2503.07536/) (2025)
- [ScienceAgentBench: Toward Rigorous Assessment of Language Agents for Data-Driven Scientific Discovery](https://mlsanity.com/paper/2410.05080/) (2024)
- [JailBreakV: A Benchmark for Assessing the Robustness of MultiModal Large Language Models against Jailbreak Attacks](https://mlsanity.com/paper/2404.03027/) (2024)
- [Mulberry: Empowering MLLM with o1-like Reasoning and Reflection via Collective Monte Carlo Tree Search](https://mlsanity.com/paper/2412.18319/) (2024)
- [Safety Fine-Tuning at (Almost) No Cost: A Baseline for Vision Large Language Models](https://mlsanity.com/paper/2402.02207/) (2024)
- [LlamaV-o1: Rethinking Step-by-step Visual Reasoning in LLMs](https://mlsanity.com/paper/2501.06186/) (2025)
- [SafeChain: Safety of Language Models with Long Chain-of-Thought Reasoning Capabilities](https://mlsanity.com/paper/2502.12025/) (2025)
- [Llama Guard 3 Vision: Safeguarding Human-AI Image Understanding Conversations](https://mlsanity.com/paper/2411.10414/) (2024)
- [Eyes Closed, Safety On: Protecting Multimodal LLMs via Image-to-Text Transformation](https://mlsanity.com/paper/2403.09572/) (2024)
- [Competitive Programming with Large Reasoning Models](https://mlsanity.com/paper/2502.06807/) (2025)
- [The Hidden Risks of Large Reasoning Models: A Safety Assessment of R1](https://mlsanity.com/paper/2502.12659/) (2025)
- [SPA-VL: A Comprehensive Safety Preference Alignment Dataset for Vision Language Model](https://mlsanity.com/paper/2406.12030/) (2024)
- [VLSBench: Unveiling Visual Leakage in Multimodal Safety](https://mlsanity.com/paper/2411.19939/) (2024)
- [Multimodal Situational Safety](https://mlsanity.com/paper/2410.06172/) (2024)
- [SafeMLRM: Demystifying Safety in Multi-modal Large Reasoning Models](https://mlsanity.com/paper/2504.08813/) (2025)
- [AStar: Boosting Multimodal Reasoning with Automated Structured Thinking](https://mlsanity.com/paper/2502.02339/) (2025)
- [Rethinking Bottlenecks in Safety Fine-Tuning of Vision Language Models](https://mlsanity.com/paper/2501.18533/) (2025)

## Built on by (most cited first, 19 of 19)

- [A Comprehensive Survey on Trustworthiness in Reasoning with Large Language Models](https://mlsanity.com/paper/2509.03871/) (2025)
- [SafeGRPO: Self-Rewarded Multimodal Safety Alignment via Rule-Governed Policy Optimization](https://mlsanity.com/paper/2511.12982/) (2025)
- [Imagination Helps Visual Reasoning, But Not Yet in Latent Space](https://mlsanity.com/paper/2602.22766/) (2026)
- [SaFeR-VLM: Toward Safety-aware Fine-grained Reasoning in Multimodal Models](https://mlsanity.com/paper/2510.06871/) (2025)
- [Phi-4-reasoning-vision-15B Technical Report](https://mlsanity.com/paper/2603.03975/) (2026)
- [Beyond Reasoning Gains: Mitigating General-Capability Forgetting in Large Reasoning Models](https://mlsanity.com/paper/2510.21978/) (2025)
- [GuardTrace-VL: Detecting Unsafe Multimodel Reasoning via Iterative Safety Supervision](https://mlsanity.com/paper/2511.20994/) (2025)
- [MTMCS-Bench: Evaluating Contextual Safety of Multimodal Large Language Models in Multi-Turn Dialogues](https://mlsanity.com/paper/2601.06757/) (2026)
- [When Helpers Become Hazards: A Benchmark for Analyzing Multimodal LLM-Powered Safety in Daily Life](https://mlsanity.com/paper/2601.04043/) (2026)
- [Safety Recovery in Reasoning Models Is Only a Few Early Steering Steps Away](https://mlsanity.com/paper/2602.11096/) (2026)
- [From Perception to Reasoning: Deep Thinking Empowers Multimodal Large Language Models](https://mlsanity.com/paper/2511.12861/) (2025)
- [Think-Reflect-Revise: A Policy-Guided Reflective Framework for Safety Alignment in Large Vision Language Models](https://mlsanity.com/paper/2512.07141/) (2025)
- [SaFeR-Steer: Evolving Multi-Turn MLLMs via Synthetic Bootstrapping and Feedback Dynamics](https://mlsanity.com/paper/2604.16358/) (2026)
- [SaFeR-ToolKit: Structured Reasoning via Virtual Tool Calling for Multimodal Safety](https://mlsanity.com/paper/2603.02635/) (2026)
- [Harnessing Textual Refusal Directions for Multimodal Safety](https://mlsanity.com/paper/2606.31876/) (2026)
- [ReFrame: Evidence-Guided Test-Time Safety Alignment in Multimodal Large Language Models](https://mlsanity.com/paper/2608.21100/) (2026)
- [ContextBias: Controlled Evaluation of Bias Persistence Under Context Shift in Text-to-Image Models](https://mlsanity.com/paper/2608.29847/) (2026)
- [A2Safe: Counterfactual Evidence-Aligned Adaptive Agent Collaboration for Safe and Effective Visual Question Answering](https://mlsanity.com/paper/2609.24098/) (2026)
- [Visual Grounding Safety in Vision-Language Models](https://mlsanity.com/paper/2610.05637/) (2026)

## Abstract (arXiv)

The rapid development of Multimodal Large Reasoning Models (MLRMs) has demonstrated broad application potential, yet their safety and reliability remain critical concerns that require systematic exploration. To address this gap, we conduct a comprehensive and systematic safety evaluation of 11 MLRMs across 5 benchmarks and unveil prevalent safety degradation phenomena in most advanced models. Moreover, our analysis reveals distinct safety patterns across different benchmarks: significant safety degradation is observed across jailbreak robustness benchmarks, whereas safety-awareness benchmarks demonstrate less pronounced degradation. In particular, the long thought process in some scenarios even enhances safety performance. Therefore, it is a potential approach to address safety issues in MLRMs by leveraging the intrinsic reasoning capabilities of the model to detect unsafe intent. To operationalize this insight, we construct a multimodal tuning dataset that incorporates a safety-oriented thought process. Experimental results from fine-tuning existing MLRMs with this dataset effectively enhances the safety on both jailbreak robustness and safety-awareness benchmarks. This study provides a new perspective for developing safe MLRMs. Our dataset is available at https://github.com/xinyuelou/Think-in-Safety.