arXiv AI

Alignment of LRMs via Counter-Aligned Few-Shot Conversation Exposure

Large Reasoning Models (LRMs) use explicit chain‑of‑thought reasoning and large context windows to perform complex tasks, but these features create new attack surfaces. The paper introduces SRCF, an attack that steers LRMs by prepending counter‑aligned few‑shot conversations with explicit CoT traces, causing unsafe outputs on harmful queries and unwarranted refusals on benign ones, without needing model internals. To counter this, the authors propose ARCF, a post‑training defense that exposes models to counter‑aligned conversational contexts while enforcing aligned targets, improving safety and helpfulness without harming utility.

arXiv Machine Learning

Sep 21

OverThink: Slowdown Attacks on Reasoning LLMs

The paper introduces OverThink, a slowdown attack that forces reasoning language models (RLMs) to produce many more reasoning tokens while still giving correct answers. By injecting decoy reasoning problems—such as Markov decision processes, language translation, or graphic comprehension—into the model’s context, attackers can dramatically increase token generation (up to 46× on SQuAD and 17× on coding agents). The study evaluates the attack on both proprietary and open-source RLMs across multiple datasets, explores multimodal and coding‑agent variants, and tests several defenses, concluding that defending against OverThink is challenging and that newer RLMs are even more vulnerable due to higher per‑token costs and increased reasoning token usage.

By Abhinav Kumar, Jaechul Roh, Ali Naseh, Marzena Karpinska, Mohit Iyyer, Amir Houmansadr, Eugene Bagdasarian

llms agents reinforcement-learning multimodal safety

arXiv AI

Jun 26

HauntAttack: When Attack Follows Reasoning as a Shadow

arXiv:2506. 07031v5 Announce Type: replace-cross Abstract: Emerging Large Reasoning Models (LRMs) consistently excel in mathematical and reasoning tasks, showcasing remarkable capabilities.

By Jingyuan Ma, Rui Li, Zheng Li, Junfeng Liu, Heming Xia, Lei Sha, Zhifang Sui

safety

arXiv AI

Jul 7

CAP-CoT: Cycle Adversarial Prompt for Improving Chain of Thoughts in LLM Reasoning

arXiv:2604. 23270v2 Announce Type: replace Abstract: Chain-of-Thought (CoT) prompting has emerged as a simple and effective way to elicit step-by-step solutions from large language models (LLMs).

By Shuxu Chen, Yitian Zhou, Jiaquan Zhang, Haoyu Bian, Wenrui Hu, Aming Wu, Sungyoung Lee, Chaoning Zhang, Hyundong Shin

llms agents benchmarks safety

arXiv AI

Sep 10

Does Deeper Reasoning Compromise Alignment? Revealing and Mitigating of Alignment Collapse in Large Reasoning Models

arXiv:2609.08186v1 Announce Type: new Abstract: The emergence of Chain-of-Thought (CoT) has established a robust foundation for Large Reasoning Models (LRMs). While deep reasoning is widely believed...

By Yu-Hang Wu, Yu-Jie Xiong, Henghua Zhang, Bairui Zhang, Jia-Chen Zhang, Shaohua Li

safety

arXiv Machine Learning

Sep 22

Assessing Adversarial Robustness of Latent Reasoning Models

arXiv:2609.22228v1 Announce Type: cross Abstract: Large language models increasingly rely on long chain-of-thought (CoT) trajectories for complex reasoning, but autoregressive generation brings subst...

By Shaolong Chen, Ang Li, Mingjie Li, Yisen Wang

llms multimodal benchmarks safety

arXiv AI

Aug 11

Dual-Adversarial Safety Alignment: Cultivating Intrinsic Threat Comprehension in LRMs

arXiv:2608. 09542v1 Announce Type: cross Abstract: Large reasoning models (LRMs) achieve remarkable success on complex tasks but remain vulnerable to harmful prompts that induce unsafe outputs.

By Hongli Shen, Shaopeng Fu, Qinbo Zhang, Jian Li, Di Wang

agents safety

arXiv AI

Jul 24

Robust Critics: Defending LLMs Against Multi-Turn Attacks

arXiv:2607. 20472v1 Announce Type: new Abstract: When a user asks a language model something harmful, is it a genuine attack or a misunderstood but well-meaning question?

By Roman Belaire, Arunesh Sinha, Pradeep Varakantham

llms reinforcement-learning fine-tuning safety

arXiv AI

Aug 6

Temporal Context Awareness: A Defense Framework Against Multi-turn Manipulation Attacks on Large Language Models

arXiv:2503. 15560v1 Announce Type: cross Abstract: Large Language Models (LLMs) are increasingly vulnerable to sophisticated multi-turn manipulation attacks, where adversaries strategically build context through seemingly benign conversational turns to circumvent safety measures and elicit harmful or unauthorized responses.

By Prashant Kulkarni, Assaf Namer

llms rag robotics safety

Hugging Face Trending Papers

Jun 13

Defending against Adaptive Prompt Injection Attacks via Reasoning-enabled Task Alignment

Indirect prompt injection attacks hijack LLM-based agents by embedding malicious instructions in third-party data that the agent retrieves during task execution. Existing defenses report near-zero attack success rate on static benchmarks, yet recent adaptive evaluations show that these results collapse once the attacker is allowed to optimize against the deployed defense.

llms rag agents reinforcement-learning multimodal benchmarks safety

arXiv Machine Learning

Aug 31

Large Reasoning Models Learn Better Alignment from Flawed Thinking

Large reasoning models generate chain-of-thought (CoT) before answering but struggle with safety alignment and can be misled by flawed premises. The paper introduces RECAP, a reinforcement learning approach that trains models to override flawed reasoning paths and produce safe, helpful responses without extra training cost. RECAP improves safety, jailbreak robustness, and reduces overrefusal while preserving core reasoning abilities and inference token budget.

By ShengYun Peng, Pin-Yu Chen, Eric Smith, Song Jiang, Hongyuan Zhan, Haozhu Wang, Mahesh Pasupuleti, Duen Horng Chau, Jianfeng Chi

reinforcement-learning safety

arXiv AI

Jun 8

Zero-Shot Embedding Drift Detection: A Lightweight Defense Against Prompt Injections in LLMs

arXiv:2601. 12359v1 Announce Type: cross Abstract: Prompt injection attacks have become an increasing vulnerability for LLM applications, where adversarial prompts exploit indirect input channels such as emails or user-generated content to circumvent alignment safeguards and induce harmful or unintended outputs.

By Anirudh Sekar, Mrinal Agarwal, Rachel Sharma, Akitsugu Tanaka, Jasmine Zhang, Arjun Damerla, Kevin Zhu

llms rag benchmarks safety

arXiv AI

Jun 16

Defending against Adaptive Prompt Injection Attacks via Reasoning-enabled Task Alignment

arXiv:2606. 15441v1 Announce Type: cross Abstract: Indirect prompt injection attacks hijack LLM-based agents by embedding malicious instructions in third-party data that the agent retrieves during task execution.

By Lipeng He, Yihan Wang, Jiawen Zhang, N. Asokan

llms rag agents reinforcement-learning multimodal benchmarks safety

Related stories