LLM Safety Alignment

LLM: Large Language Model

Follow this topic Ask about this topic

Momentum

25 papers in the last four weeks, up 14% on the four weeks before. 0.2% of all new papers.

Jul 13Week of Sep 28

Latest papers 236

All topics

When Choices Become Risks: Safety Failures of Large Language Models under Multiple-Choice Constraints
Apr 18, 2026Yuheng Chen, Zhiyu Wu, Bowen Cheng +2Adversarial Attacks on LLMs LLM Safety Alignment
Kagoshima University · Fudan University · China University of Petroleum-Beijing
DART: Mitigating Harm Drift in Difference-Aware LLMs via Distill-Audit-Repair Training
Apr 18, 2026Ziwen Pan, Zihan Liang, Jad Kabbara +1LLM Auditing Social Bias in Language Models
Emory University · MIT
Benign Fine-Tuning Breaks Safety Alignment in Audio LLMs
Apr 17, 2026Jaechul Roh, Virat Shejwalkar, Amir HoumansadrLanguage Model Safety Evaluation LLM Fine-Tuning
University of Massachusetts Amherst
From Consumption to Reflection: Designing Human-AI Relations for Stable Reasoning
Apr 17, 2026Rikard Rosenbacke, Carl Rosenbacke, Victor Rosenbacke +1LLM Safety Alignment Human Oversight of AI
Faculty of Medicine, Lund University, Sweden · Department of Economics, Lund University School of Economics and Management, Sweden · Department of Health Services Research and Policy, London School of Hygiene & Tropical Medicine, UK
Pruning Unsafe Tickets: A Resource-Efficient Framework for Safer and More Robust LLMs
Apr 17, 2026Wai Man Si, Mingjie Li, Michael Backes +1LLM Pruning LLM Safety Alignment
CISPA Helmholtz Center for Information Security
Safety Training Modulates Harmful Misalignment Under On-Policy RL, But Direction Depends on Environment Design
Apr 14, 2026Leon Eshuijs, Shihan Wang, Antske FokkensReward Hacking RL for Language Models
Vrije Universiteit Amsterdam · Utrecht University
Activation Steering for Aligned Open-ended Generation without Sacrificing Coherence
Apr 9, 2026Niklas Herbster, Martin Zborowski, Alberto Tosato +2LLM Safety Alignment Language Model Robustness
1Tara Research · 2Technical University of Munich · 3Mila Quebec AI Institute
SafeMath: Safe Solutions for Unsafe Math Word Problems
Mar 26, 2026Sagnik Basu, Subhrajit Mitra, Aman Juneja +3LLM Safety Alignment LLM Safety Evaluation
1Indian Institute of Technology Kharagpur · 2Cisco Systems · 3National University of Singapore
Safe RLHF Beyond Expectation: Stochastic Dominance for Universal Spectral Risk Control
Mar 11, 2026Yaswanth Chittepu, Ativ Joshi, Rajarshi Bhattacharjee +1Constrained RL LLM Safety Alignment
University of Massachusetts Amherst
Mitigating the Safety-utility Trade-off in LLM Alignment via Adaptive Safe Context Learning
Feb 14, 2026Yanbo Wang, Minzheng Wang, Jian Liang +3LLM Alignment RL for Language Model Reasoning
School of Artificial Intelligence, University of Chinese Academy of Sciences, Beijing, China. · NLPR & MAIS, Institute of Automation, Chinese Academy of Sciences, Beijing, China. · Ritzz-AI.
Towards Multi-modal Multi-turn Safety: From Agentic Interaction to Strategic Alignment
Jan 8, 2026Han Zhu, Jiale Chen, Chengkun Cai +8Language Model Safety Evaluation Multimodal Large Language Models
Hong Kong University of Science and Technology · University College London · AISpeech +1
Agent Tools Orchestration Leaks More: Dataset, Benchmark, and Mitigation
Dec 18, 2025Yuxuan Qiao, Dongqin Liu, Hongchang Yang +2LLM Agent Security LLM Agents
Institute of Information Engineering, Chinese Academy of Sciences · School of Cyber Security, University of Chinese Academy of Sciences
Response-Based Knowledge Distillation for Multilingual Jailbreak Prevention Unwittingly Compromises Safety
Dec 8, 2025Max Zhang, Derek Liu, Kai Zhang +2Adversarial Attacks on LLMs LLM Safety Alignment
Reasoning Up the Instruction Ladder for Controllable Language Models
Oct 30, 2025Zishuo Zheng, Vidhisha Balachandran, Chan Young Park +2LLM Safety Alignment Instruction Following
Department of Computer Science and Engineering, The Ohio State University · Microsoft Research · Allen Institute for AI
Attention Misses Visual Risk: Risk-Adaptive Steering for Multimodal Safety Alignment
Oct 15, 2025Jonghyun Park, Minhyuk Seo, Chaewon Yeo +1Multimodal Robustness LLM Safety Alignment
Seoul National University · KU Leuven
The Alignment Auditor: A Bayesian Framework for Verifying and Refining LLM Objectives
Oct 7, 2025Matthieu Bou, Nyal Patel, Arjun Jagota +2Reinforcement Learning LLM Auditing
1Imperial College London · 2Amazon AGI
Turning the Spell Around: Lightweight Alignment Amplification via Rank-One Safety Injection
Aug 28, 2025Harethah Abu Shairah, Hasan Abed Al Kader Hammoud, George Turkiyyah +1LLM Guardrails LLM Safety Alignment
King Abdullah University of Science and Technology (KAUST)
Context Misleads LLMs: The Role of Context Filtering in Maintaining Safe Alignment of LLMs
Aug 9, 2025Jinhwa Kim, Ian G. HarrisLLM Safety Alignment LLM Safety
Department of Computer Science University of California, Irvine
Chasing Moving Targets with Online Self-Play Reinforcement Learning for Safer Language Models
Jun 9, 2025Mickel Liu, Liwei Jiang, Yancheng Liang +4Self-Play RL Adversarial Attacks on LLMs
Paul G. Allen School of Computer Science & Engineering, University of Washington, Seattle, WA, USA · Department of Computer Science, Stanford University, Stanford, CA, USA
Decoding One Safety Trigger Token for Balancing Safety and Usability in Large Language Models
May 12, 2025Haoran Gu, Handing Wang, Yi Mei +2LLM Safety Alignment LLM Safety
Xidian University · Victoria University of Wellington · Westlake University
Adaptive Helpfulness-Harmlessness Alignment with Preference Vectors
Apr 27, 2025Ren-Wei Liang, Chin-Ting Hsu, Chan-Hung Yu +6LLM Alignment LLM Safety Alignment
National Taiwan University · Appier AI Research · Graduate Institute of Communication Engineering, National Taiwan University +1
Decoupled Alignment for Robust Plug-and-Play Adaptation
Jun 3, 2024Haozheng Luo, Jiahao Yu, Wenxin Zhang +9LLM Alignment LLM Safety Alignment
Northwestern University · New York University Abu Dhabi · Stanford University +3
Learning diverse attacks on large language models for robust red-teaming and safety tuning
May 28, 2024Seanie Lee, Minsu Kim, Lynn Cherif +8Adversarial Attacks on LLMs LLM Red Teaming
KAIST · Mila – Québec AI Institute · Université de Montréal +3
Eraser: Jailbreaking Defense in Large Language Models via Unlearning Harmful Knowledge
Apr 8, 2024Weikai Lu, Ziqian Zeng, Jianwei Wang +5LLM Safety Alignment LLM Jailbreak Attacks
South China University of Technology, China · Pazhou Laboratory, China
SaFeR-Steer: Evolving Multi-Turn MLLMs via Synthetic Bootstrapping and Feedback Dynamics
Date pendingHaolong Hu, Hanyu Li, Tiancheng He +6Language Model Steering Multimodal Large Language Models
Huazhong University of Science and Technology · Beijing University of Posts and Telecommunications · West China Biomedical Big Data Center, Sichuan University +2
Norms at a Price: Why RL-Based Alignment Can Promise Conditional Compliance at Best
Date pendingKevin Baum, Rūta Binkytė, Felix JahnAI Alignment LLM Safety Alignment
German Research Center for Artificial Intelligence (DFKI), Saarbrücken · Institute for Ethics in Technology, Hamburg University of Technology (TUHH) · Oxford Internet Institute, University of Oxford

Apr 18, 2026, cs.CLJ/K move · Enter open · S save

When Choices Become Risks: Safety Failures of Large Language Models under Multiple-Choice Constraints

Yuheng Chen, Zhiyu Wu, Bowen Cheng, Yihang Wu+1

Kagoshima University · Fudan University · China University of Petroleum-Beijing

Adversarial Attacks on LLMs LLM Safety Alignment LLM Safety Evaluation Refusal Behavior in Language Models

Open paper AskPDF arXiv

We identify and systematically characterize a class of task-structural alignment failures in large language models (LLMs): even when the harmful intent remains unchanged, changing the task presentation and output constraints can substantially alter model safety behavior. Specifically, when a harmful request is reformulated as a forced-choice multiple-choice question (MCQ) in which all options are harmful and no refusal option is provided, some models that refuse the equivalent open-ended query instead select, prefer, or justify a harmful option. We evaluate 14 proprietary and open-source models on a bilingual Chinese-English human-authored dataset covering five harm categories, together with 900 model-generated Chinese adversarial MCQs. On human-authored data, attack success rate (ASR) increases sharply as prompts shift from open-ended queries to explicit forced-choice formats, typically peaking under intermediate levels of choice constraint. Model-generated Chinese MCQs further weaken or eliminate the recovery regime observed on human-authored data, driving ASR close to saturation for multiple models. The observed transfer patterns are consistent with stronger generators producing more difficult or boundary-adjacent MCQs, although other properties of the generated inputs may also contribute. We also find that adding an explicit refusal option or a safety preamble substantially reduces ASR for several high-capability models, often to near-zero levels, although their effectiveness varies across target models. These findings suggest that safety evaluations centered on open-ended generation may underestimate risks in structured deployment settings, and that task structure should be treated as an important and diagnosable dimension of safety evaluation and alignment training.

Momentum

Related topics

Latest papers 236