mlsanityData to 14 Oct 2026

Conflicts Make Large Reasoning Models Vulnerable to Attacks

arXiv 2604.09750 ↗First posted 10 Apr 2026cs.CR cs.AIcited by 1

On the map: LLM reasoning → Reasoning for agents, planning and code

Abstract (arXiv)

Large Reasoning Models (LRMs) have achieved remarkable performance across diverse domains, yet their decision-making under conflicting objectives remains insufficiently understood. This work investigates how LRMs respond to harmful queries when confronted with two categories of conflicts: internal conflicts that pit alignment values against each other and dilemmas, which impose mutually contradictory choices, including sacrificial, duress, agent-centered, and social forms. Using over 1,300 prompts across five benchmarks, we evaluate three representative LRMs - Llama-3.1-Nemotron-8B, QwQ-32B, and DeepSeek R1 - and find that conflicts significantly increase attack success rates, even under single-round non-narrative queries without sophisticated auto-attack techniques. Our findings reveal through layerwise and neuron-level analyses that safety-related and functional representations shift and overlap under conflict, interfering with safety-aligned behavior. This study highlights the need for deeper alignment strategies to ensure the robustness and trustworthiness of next-generation reasoning models. Our code is available at https://github.com/DataArcTech/ConflictHarm. Warning: This paper contains inappropriate, offensive and harmful content.

Key facts. Cited by 1 paper 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 35 references that are in the map. First posted on arXiv 10 Apr 2026.

Builds on

33 of 35 references.

Closest ancestors · 4

Same direction, newest first

  1. 2025A Mousetrap: Fooling Large Reasoning Models for Jailbreak with Chain of Iterative Chaos19 Feb
  2. 2025H-CoT: Hijacking the Chain-of-Thought Safety Reasoning Mechanism to Jailbreak Large Reasoning Models, Including OpenAI o1/o3, DeepSeek-R1, and Gemini 2.0 Flash Thinking18 Feb
  3. 2025SafeChain: Safety of Language Models with Long Chain-of-Thought Reasoning Capabilities17 Feb
  4. 2024How Johnny Can Persuade LLMs to Jailbreak Them: Rethinking Persuasion to Challenge AI Safety by Humanizing LLMs12 Jan

Other references · 29

  1. 2024DeepSeekMath: Pushing the Limits of Mathematical Reasoning in Open Language Modelscited by 9,379
  2. 2023Universal and Transferable Adversarial Attacks on Aligned Language Modelscited by 3,104
  3. 2023Jailbreaking Black Box Large Language Models in Twenty Queriescited by 1,420
  4. 2023Llama Guard: LLM-based Input-Output Safeguard for Human-AI Conversationscited by 1,207
  5. 2023Fine-tuning Aligned Language Models Compromises Safety, Even When Users Do Not Intend To!cited by 1,154
  6. 2023Large Language Modelscited by 780
  7. 2023Catastrophic Jailbreak of Open-source LLMs via Exploiting Generationcited by 451
  8. 2023DeepInception: Hypnotize Large Language Model to Be Jailbreakercited by 367
  9. 2023Toxicity in ChatGPT: Analyzing Persona-assigned Language Modelscited by 360
  10. 2023Jailbreak in pieces: Compositional Adversarial Attacks on Multi-Modal Language Modelscited by 273
  11. 2024Assessing the Brittleness of Safety Alignment via Pruning and Low-Rank Modificationscited by 231
  12. 2023Evaluating the Moral Beliefs Encoded in LLMscited by 215
17 more
  1. 2023Scalable and Transferable Black-Box Jailbreaks for Language Models via Persona Modulationcited by 214
  2. 2022On Second Thought, Let's Not Think Step by Step! Bias and Toxicity in Zero-Shot Reasoningcited by 194
  3. 2024Free Process Rewards without Process Labelscited by 165
  4. 2024Safety Layers in Aligned Large Language Models: The Key to LLM Securitycited by 134
  5. 2024Forest-of-Thought: Scaling Test-Time Compute for Enhancing LLM Reasoningcited by 110
  6. 2024The Benefits of a Concise Chain of Thought on Problem-Solving in Large Language Modelscited by 63
  7. 2024Language Model Alignment in Multilingual Trolley Problemscited by 46
  8. 2023Probing the Moral Development of Large Language Models through Defining Issues Testcited by 25
  9. 2025Will Systems of LLM Agents Cooperate: An Investigation into a Social Dilemmacited by 25
  10. 2025Mix Data or Merge Models? Balancing the Helpfulness, Honesty, and Harmlessness of Large Language Model via Model Mergingcited by 23
  11. 2024Are Large Language Models More Empathetic than Humans?cited by 18
  12. 2025LLMs are Vulnerable to Malicious Prompts Disguised as Scientific Languagecited by 17
  13. 2024Exploring and steering the moral compass of Large Language Modelscited by 16
  14. 2025Bullying the Machine: How Personas Increase LLM Vulnerabilitycited by 9
  15. 2025Normative Conflicts and Shallow AI Alignmentcited by 5
  16. 2025Survival at Any Cost? LLMs and the Choice Between Self-Preservation and Human Harmcited by 3
  17. 2025CASTILLO: Characterizing Response Length Distributions of Large Language Modelscited by 1

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 1.

1 paper.

  1. 2026Alignment of LRMs via Counter-Aligned Few-Shot Conversation Exposurecited by 0

Similar, not cited

Close in wording, no citation either way

  1. 2026On the Efficiency-Safety Dilemma in Large Reasoning Models
  2. 2026Consistency of Large Reasoning Models Under Multi-Turn Attacks
  3. 2026Reasoning Structure Matters for Safety Alignment of Reasoning Models
  4. 2026Reasoning Hijacking: The Fragility of Reasoning Alignment in Large Language Models
  5. 2025RAJ-PGA: Reasoning-Activated Jailbreak and Principle-Guided Alignment Framework for Large Reasoning Models
  6. 2026MADBench: Benchmarking the Security of Multi-Agent Debate
  7. 2025SafeRBench: Dissecting the Reasoning Safety of Large Language Models
  8. 2025Beyond Safe Answers: A Benchmark for Evaluating True Risk Awareness in Large Reasoning Models