Large Language Model Safety

Follow this field Ask about this field

Recent momentum

emerging

0 papers in the last 28 days · 0.0% of indexed attention

Twelve weeks of publication activity for this topic as it is defined today.

Weekly history

Recent digests

What was published in this field, kept on the site without email delivery.

Period ending 2026-09-21

27 new papers

A weekly snapshot of new work published in Large Language Model Safety.

Period ending 2026-09-14

12 new papers

A weekly snapshot of new work published in Large Language Model Safety.

Inside this field

Focused directions

Focused topic

Large Language Model Safety

Within Large Language Model Safety

532 papersExplore

Focused topic

Language Model Evasion Attacks

Within Large Language Model Safety

204 papersExplore

Focused topic

Adversarial Evaluation

Within Large Language Model Safety

114 papersExplore

Focused topic

Adversarial Prompting

Within Large Language Model Safety

30 papersExplore

855 papers

Latest in Large Language Model Safety

Browse all topics →

RefusalBench: Why Refusal Rate Misranks Frontier LLMs on Biological Research Prompts
May 20, 2026Lukas Weidener, Marko Brkić, Mihailo Jovanović +2Refusals Frontier Models
Hack-Verifiable Environments: Towards Evaluating Reward Hacking at Scale
May 20, 2026Amit Roth, Ankur Samanta, Matan Halevy +2Hacking Outcome-Based Rewards
An Application-Layer Multi-Modal Covert-Channel Reference Monitor for LLM Agent Egress
May 20, 2026Alfredo MetereLanguage Model Evasion Attacks Leakage
Trusted Weights, Treacherous Optimizations? Optimization-Triggered Backdoor Attacks on LLMs
May 20, 2026Yifei Wang, Tianlin Li, Xiaohan Zhang +3Language Model Evasion Attacks Backdoor Attacks
QwenSafe: Multimodal Content Rating Description Identification via Preference-Aligned VLMs
May 20, 2026Dishanika Denipitiyage, Aruna Seneviratne, Suranga SeneviratneMultimodal Evaluation Content Moderation
Measuring Safety Alignment Effects in Autonomous Security Agents
May 19, 2026Isaac David, Arthur GervaisLarge Language Model Safety Safety Alignment
The Evaluation Game: Beyond Static LLM Benchmarking
May 19, 2026Paul Wang, Jade Garcia-Bourrée, Anne-Marie Kermarrec +1Adversarial Evaluation Benchmarking Large Language Models
Exploring and Developing a Pre-Model Safeguard with Draft Models
May 19, 2026Hongyu Cai, Arjun Arunasalam, Yiming Liang +2Large Language Model Safety Safeguard
DECOR: Auditing LLM Deception via Information Manipulation Theory
May 19, 2026Linyue Cai, Samuel Yeh, Jwala Dhamala +2Deception Language Model Evasion Attacks
Conformal Selective Acting: Anytime-Valid Risk Control for RLVR-Trained LLMs
May 18, 2026Hamed Khosravi, Xiaoming HuoConformal Risk Control Reinforcement Learning With Verifiable Reward
Be Kind, Rewrite: Benign Projections via Rewriting Defend Against LLM Data Poisoning Attacks
May 18, 2026John T. Halloran, Noopur S. BhattLanguage Model Evasion Attacks Data Poisoning
POLAR-Bench: A Diagnostic Benchmark for Privacy-Utility Trade-offs in LLM Agents
May 18, 2026Qiaoyuan Zheng, Yiqu Yang, Qi Gao +1Adversarial Evaluation Privacy-Utility Trade-Offs
A Survey of Large Audio Language Models: Generalization, Trustworthiness, and Outlook
May 18, 2026Kaiwen Luo, Zhenhong Zhou, Leo Wang +31Large Audio Language Models Large Language Model Safety
Position: A Three-Layer Probabilistic Assume-Guarantee Architecture Is Structurally Required for Safe LLM Agent Deployment
May 18, 2026S. Bensalem, Y. Dong, M. Franzle +6Large Language Model Safety Large Language Model Agents
OEP: Poisoning Self-Evolving LLM Agents via Locally Correct but Non-Transferable Experiences
May 18, 2026Kaixiang Wang, Jiong Lou, Zhaojiacheng Zhou +1Language Model Evasion Attacks Poisoning
It Takes Two: Complementary Self-Distillation for Contextual Integrity in LLMs
May 18, 2026Sangwoo Park, Woongyeong Yeo, Seanie Lee +6Privacy-Utility Trade-Offs Large Language Model Safety
REBAR: Reference Ethical Benchmark for Autonomy Readiness
May 18, 2026Jonathan Diller, David Barnes, Rebekah Bogdanoff +14Autonomy Adversarial Evaluation
Acoustic Interference: A New Paradigm Weaponizing Acoustic Latent Semantic for Universal Jailbreak against Large Audio Language Models
May 18, 2026Yanyun Wang, Yu Huang, Zi Liang +2Large Audio Language Models Large Language Model Safety
An Empirical Study of Privacy Leakage Chains via Prompt Injection in Black-Box Chatbot Environments
May 18, 2026Hongjang Yang, Hyunsik Na, Daeseon ChoiIndirect Prompt Injection Language Model Evasion Attacks
Safety Geometry Collapse in Multimodal LLMs and Adaptive Drift Correction
May 18, 2026Jiahe Guo, Xiangran Guo, Jiaxuan Chen +6Large Language Model Safety Multi-Modal Large Language Models
Remembering More, Risking More: Longitudinal Safety Risks in Memory-Equipped LLM Agents
May 18, 2026Ahmad Al-Tawaha, Shangding Gu, Peizhi Niu +2Agent Memory Large Language Model Safety
Geometry-Lite: Interpretable Safety Probing via Layer-Wise Margin Geometry
May 18, 2026Woo Seob Sim, Yu Rang ParkLarge Language Model Safety Bias Learning
SafeLens: Deliberate and Efficient Video Guardrails with Fast-and-Slow Screening
May 17, 2026Shahriar Kabir Nahin, Hadi Askari, Muhao Chen +1Streaming Guardrails Inference-Time Defense
Adaptive Generate-Rank-Verify: Inference-Time Search with Costly Verification
May 17, 2026Shaddin Dughmi, Mahdi Haghifam, Yusuf Hakan KalayciInference-Time Verifier
Trust No Tool: Evaluating and Defending LLM Agents under Untrusted Tool Feedback
May 17, 2026Lecheng Yan, Ruizhe Li, Xicheng Han +5Untrusted Input Large Language Model Safety
Ablating Safety: Mechanisms for Removing Alignment in Language Models for Security Applications
May 17, 2026Isaac David, Arthur GervaisLarge Language Model Safety Value Alignment
Towards trustworthy agentic AI: a comprehensive survey of safety, robustness, privacy, and system security
May 17, 2026Jinhu Qi, Muzhi Li, Jiahong Liu +9Trustworthy Artificial Intelligence Agentic Systems
LPG: Balancing Efficiency and Policy Reasoning in Latent Policy Guardrails
May 17, 2026Nanxi Li, Zhengyue Zhao, Chaowei XiaoStreaming Guardrails Large Language Model Safety
When Efficiency Backfires: Cascading LLMs Trigger Cascade Failure under Adversarial Attack
May 17, 2026Zehan Sun, Dingfan Chen, Songze LiLanguage Model Evasion Attacks Large Language Model Performance
Why Do Safety Guardrails Degrade Across Languages?
May 16, 2026Max Zhang, Ameen Patel, Sang T. Truong +1Multilingual Safety Large Language Model Safety
DriveSafe: A Framework for Risk Detection and Safety Suggestions in Driving Scenarios
May 16, 2026Sainithin Artham, Shankar Gangisetty, Avijit Dasgupta +1Autonomous Driving Drives
Multi-Paradigm Agent Interaction in Practice:A Systematic Analysis of Generator-Evaluator, ReAct Loop,and Adversarial Evaluation in the buddyMe Framework
May 16, 2026Xiaohua Wang, Chao Han, Kai Yu +2Multi-Agent Systems Adversarial Evaluation
Distinguishable Deletion: Unifying Knowledge Erasure and Refusal for Large Language Model Unlearning
May 16, 2026Puning Yang, Junchi Yu, Qizhou Wang +3Large Language Model Unlearning Unlearning Aims
ZeroUnlearn: Few-Shot Knowledge Unlearning in Large Language Models
May 16, 2026Yujie Lin, Chengyi Yang, Zhishang Xiang +2Large Language Model Unlearning Machine Unlearning
State Contamination in Memory-Augmented LLM Agents
May 16, 2026Yian Wang, Agam Goyal, Yuen Chen +1Agent Memory Large Language Model Agents
Formal Methods Meet LLMs: Auditing, Monitoring, and Intervention for Compliance of Advanced AI Systems
May 15, 2026Parand A. Alamdari, Toryn Q. Klassen, Sheila A. McIlraithArtificial Intelligence Governance Large Language Model Safety
A Unified Perturbation Framework for Analyzing Leaderboard Stability and Manipulation
May 15, 2026Hosna Oyarhoseini, Jimmy Lin, Amir-Hossein KarimiLeaderboard Perturbations
Asking Back: Interaction-Layer Antidistillation Watermarks
May 15, 2026Guang Yang, Amir Ghasemian, Fengchen Liu +3Language Model Evasion Attacks Watermarks
parallelcbf: A composable safety-filter and auditability framework for tensor-parallel reinforcement learning
May 15, 2026Yijun Lu, Zilei Yang, Yuyin MaControl Barrier Functions Adversarial Evaluation
Muse Spark Safety & Preparedness Report
May 14, 2026Cristina Menghini, Peter Ney, Hamza Kwisaba +117Large Language Model Safety Threat Models
Ensemble Monitoring for AI Control: Diverse Signals Outweigh More Compute
May 14, 2026Eugene Koran, Yejun Yun, Samantha Tetef +2Artificial Intelligence Safety Adversarial Evaluation
MetaBackdoor: Exploiting Positional Encoding as a Backdoor Attack Surface in LLMs
May 14, 2026Rui Wen, Mark Russinovich, Andrew Paverd +2Backdoor Attacks Language Model Evasion Attacks
Widening the Gap: Exploiting LLM Quantization via Outlier Injection
May 14, 2026Xiaohua Zhan, Kazuki Egashira, Robin Staab +2Large Language Model Quantization Quantization
Talk is (Not) Cheap: A Taxonomy and Benchmark Coverage Audit for LLM Attacks
May 14, 2026Karthik Raghu Iyer, Yazdan Jamshidi, Nicholas Bray +1Language Model Evasion Attacks Security Analysis
IFPV: An Integrated Multi-Agent Framework for Generative Operational Planning and High-Fidelity Plan Verification
May 14, 2026Zhigao Huang, Zhengqing Hu, Dong Chen +5Large Language Model Planning Adversarial Evaluation
Known By Their Actions: Fingerprinting LLM Browser Agents via UI Traces
May 14, 2026William Lugoloobi, Samuelle Marro, Jabez Magomere +2Language Model Evasion Attacks Web Agents
Selective Safety Steering via Value-Filtered Decoding
May 14, 2026Bat-Sheva Einbinder, Hen Davidov, Yee Whye Teh +2Large Language Model Safety Safety
Reducing the Safety Tax in LLM Safety Alignment with On-Policy Self-Distillation
May 14, 2026Yu Fu, Longxuan Yu, Haz Sameen Shahgir +4Unsupervised On-Policy Self-Distillation Safety Alignment
Watermarking Game-Playing Agents in Perfect-Information Extensive-Form Games
May 14, 2026Juho Kim, Fei Fang, Tuomas SandholmLarge Language Model Watermarks Watermarking
Paraphrasing Attack Resilience of Various AI-Generated Text Detection Methods
May 14, 2026Andrii Shportko, Inessa VerbitskyMachine-Generated Text Detection Artificial Intelligence Detection
GradShield: Alignment Preserving Finetuning
May 13, 2026Zhanhao Hu, Xiao Huang, Patrick Mendoza +4Large Language Model Safety Large Language Model Alignment
ExploitBench: A Capability Ladder Benchmark for LLM Cybersecurity Agents
May 13, 2026Seunghyun Lee, David BrumleyExploitation Vulnerable Code
ROK-FORTRESS: Measuring the Effect of Geopolitical Transcreation for National Security and Public Safety
May 13, 2026Michael S. Lee, Yash Maurya, Drew Rein +13Multilingual Safety Large Language Model Safety
Measuring and Mitigating Toxicity in Large Language Models: A Comprehensive Replication Study
May 13, 2026Mokshit Surana, Archit Rathod, Akshaj SatishkumarLarge Language Model Safety Toxicity
VectraYX-Nano: A 42M-Parameter Spanish Cybersecurity Language Model with Curriculum Learning and Native Tool Use
May 13, 2026Juan S. SantillanaLarge Language Model Safety Cybersecurity
Identifying AI Web Scrapers Using Canary Tokens
May 13, 2026Steven Seiden, Triss Ren, Caroline Zhang +3Online Data Selection Language Model Evasion Attacks
AERIC: Anticipatory Hidden-State Monitoring for Implicit Harmful Dialogue
May 13, 2026Jihyung Park, Saleh Afroogh, Junfeng JiaoLarge Language Model Safety Trigger
Model-Agnostic Lifelong LLM Safety via Externalized Attack-Defense Co-Evolution
May 13, 2026Xiaozhe Zhang, Chaozhuo Li, Hui Liu +4Large Language Model Safety Model-Agnostic Defense
Inducing Overthink: Hierarchical Genetic Algorithm-based DoS Attack on Black-Box Large Language Reasoning Models
May 13, 2026Shuqiang Wang, Wei Cao, Jiaqi Weng +4Large Language Model Reasoning Large Reasoning Models
Tracing Persona Vectors Through LLM Pretraining
May 13, 2026Viktor Moskvoretskii, Dominik Glandorf, Jorge Medina Moreira +2Personality Large Language Model Safety

May 20, 2026 · cs.SEJ/K move · Enter open · S save

RefusalBench: Why Refusal Rate Misranks Frontier LLMs on Biological Research Prompts

Lukas Weidener, Marko Brkić, Mihailo Jovanović, Emre Ulgac+1

Applied Scientific Intelligence, Inc.

Frontier large language models are increasingly deployed as orchestration backbones for biological research workflows, yet no shared evidence base exists for comparing their refusal behaviour on legitimate research prompts. RefusalBench, introduced here, is a matched-triple benchmark of 141 prompts in 47 bundles that holds task framing constant while varying only biological risk tier (benign, borderline, dual-use), enabling tier-conditioned comparisons robust to subdomain confounding. A 15-prompt should-refuse positive-control module establishes per-model calibration floors; three models fail to refuse even these prompts. Across 19 frontier models in the May 2026 snapshot, strict refusal rates span 0.1% to 94.6% on identical prompts. Jurisdiction does not predict refusal in this snapshot (Mann-Whitney U, p = 0.393; EU n = 1, US bimodal); provider identity does, with Anthropic's API stack predicting refusal at OR = 21.03 (95% CI: 14.58-30.34 prompt-clustered; 5.70-77.55 under model-clustered GEE). This effect is best read as access-path-level rather than model-weight-level: 99.8% of Anthropic's strict refusals carry the same safety_policy adjudicated reason code, consistent with a small set of canonical refusal templates rather than case-by-case model reasoning. Strict refusal rate misranks safety calibration: Grok 4.20 achieves the highest tier discrimination (Youden's J = 0.787) while ranking only seventh by overall refusal rate, and Claude Opus 4.7's J dropped 65% from prior versions with no improvement in dual-use detection. Nine of 18 frontier models exhibit a hedge-but-help partial-compliance pattern at dual-use tier that binary refusal metrics cannot detect.

Refusals Frontier Models Large Language Model Safety Life Sciences Research Prompt Engineering

Open paper Ask about it PDF