LLM Safety

LLM: Large Language Model

Follow this topic Ask about this topic

Momentum

36 papers in the last four weeks, up 200% on the four weeks before. 0.4% of all new papers.

Jul 13Week of Sep 28

Latest papers 272

All topics

Opir: Efficient Multi-Task Safety Classification for Toxicity, Jailbreaks, Hate Speech, and Harmful Content
May 28, 2026Ihor Stepanov, Aleksandr SmechovLanguage Model Safety Evaluation LLM Guardrails
1Knowledgator, Kyiv, Ukraine · 2Wordcab, New York, USA
Inform, Coach, Relate, Listen: Auditing LLM Caregiving Support Roles
May 28, 2026Drishti Goel, Agam Goyal, Veda Duddu +8Healthcare LLM Auditing
University of Illinois Urbana-Champaign · University of Massachusetts Amherst · OSF HealthCare +1
Robust and Efficient Guardrails with Latent Reasoning
May 27, 2026Siddharth Sai, Xiaofei Wen, Muhao ChenLLM Guardrails Content Moderation
University of California, Davis
Training Stratigraphy: Persistent Behavioral Artifacts in Large Language Models Observed Through Longitudinal AI-Human Interaction
May 27, 2026Chen Ying Claude, Zhihan LuoLLM Auditing Human-in-the-Loop AI
*Anthropic · †Independent Researcher
Where Does Toxicity Live? Mechanistic Localization and Targeted Suppression in Language Models
May 27, 2026Himanshu Beniwal, Mayank SinghLanguage Model Safety Evaluation LLM Interpretability
Indian Institute of Technology Gandhinagar
Evolving and Detecting Multi-Turn Deception using Geometric Signatures
May 26, 2026Surender Suresh Kumar, Mary L. CummingsAdversarial Prompt Generation Deception in Language Models
George Mason University Fairfax, USA
BAIT: Boundary-Guided Disclosure Escalation LLM Jailbreaking via Self-Conditioned Reasoning
May 26, 2026Xuan Luo, Yue Wang, Geng Tu +2Jailbreak Attacks LLM Safety
The Harbin Institute of Technology, Shenzhen, China · The Hong Kong Polytechnic University, Hong Kong, China · Shenzhen University, Shenzhen, China +1
ReasonOps: A Unified Operational Paradigm for Trustworthy Verified LLM Reasoning
May 26, 2026Adnan RashidLLM Safety LLM Reasoning
School of Electrical Engineering and Computer Science (SEECS) National University of Sciences and Technology (NUST) Islamabad, Pakistan
Conceptual Steganography
May 26, 2026Zhejian Zhou, Jonathan MayCoT Reasoning LLM Security
University of Southern California · Information Sciences Institute
Open-Weight LLM Fine-Tuning Defenses are Susceptible to Simple Attacks
May 26, 2026Kevin Kuo, Virginia Smith, Chhavi YadavLLM Fine-Tuning Adversarial Attacks on LLMs
Carnegie Mellon University · Simons Institute, UC Berkeley
Model Unlearning Objectives Vary for Distinct Language Functions
May 26, 2026Berk Atil, Vipul Gupta, Rebecca J. PassonneauLLM Safety Machine Unlearning
Pennsylvania State University · Scale AI
$D^2$ -Monitor: Dynamic Safety Monitoring for Diffusion LLMs via Hesitation Signals
May 25, 2026Aoxi Liu, Yupeng Chen, James Oldfield +5Language Model Safety Evaluation LLM Safety
Torr Vision Group, University of Oxford · The Chinese University of Hong Kong, Shenzhen
Preferred, Not Safer: Pairwise Preference Is a Poor Proxy for Clinical Safety
May 25, 2026Fay Elhassan, David Sasu, Alexandra Kulinkina +2LLM Safety Benchmarks Human Preference Evaluation
LiGHT Laboratory, EPFL, Lausanne, Switzerland
StructBreak: Structural Cognitive Overload-Induced Safety Failures in MLLMs
May 25, 2026Yang Luo, Xinran Liu, Tiantian Ji +3Adversarial Attacks on VLMs Language Model Safety Evaluation
Key Laboratory of Trustworthy Distributed Computing and Service (MoE), Beijing University of Posts and Telecommunications · Institute of Computing Technology, Chinese Academy of Sciences
The Age of Curiosity Meets the Age of AI: Benchmarking Child Safety in Large Language Models
May 25, 2026Samee Arif, Angana Borah, Rada MihalceaLLM Safety Benchmarks Language Model Safety Evaluation
University of Michigan
Furina: Fragmented Uncertainty-Driven Refusal Instability Attack
May 24, 2026Tongxi Wu, Jian Zhang, Yang GaoAdversarial Attacks on VLMs LLM Safety
School of Intelligence Science and Technology, Nanjing University, Suzhou, China · State Key Laboratory for Novel Software Technology, Nanjing University, Nanjing, China
Inverting the Shield: Systematically Generating Safety Tests from Policy Specifications
May 24, 2026Xiaoyue Lu, Xianglin Yang, Haijun Liu +4Automated Software Testing Language Model Safety Evaluation
Shenzhen Campus of Sun Yat-sen University, Shenzhen, China · National University of Singapore, Singapore · Independent Researcher
Jailbreak to Protect: Buffering and Reinforcing via Temporary Jailbreaking for Safe Fine-Tuning in Large Language Models
May 23, 2026Seokil Ham, Jaehyuk Jang, Wonjun Lee +1LLM Fine-Tuning LLM Safety Alignment
School of Electrical Engineering, Korea Advanced Institute of Science and Technology (KAIST), Daejeon, Republic of Korea.
JT-SAFE-V2: Safety-by-Design Foundation Model with World-Context Data
May 23, 2026Junlan Feng, Fanyu Meng, Chong Long +12LLM Safety Alignment LLM Safety
JIUTIAN Research
Safety-Oriented Routing Analysis of Mixtral MoE Under Benign and Harmful Prompts
May 22, 2026Md Nurul Absar SiddikyMixture-of-Experts Language Models Expert Routing
Department of Electrical and Computer Engineering University of Hawai’i at M¯anoa Honolulu, HI, USA
Test-Time Training Undermines Safety Guardrails
May 21, 2026Simone Antonelli, Sadegh Akhondzadeh, Aleksandar BojchevskiLanguage Model Safety Evaluation Adversarial Attacks on LLMs
CISPA Helmholtz Center for Information Security · University of Cologne
Latent-space Attacks for Refusal Evasion in Language Models
May 20, 2026Giorgio Piras, Raffaele Mura, Fabio Brau +4Adversarial Attacks on LLMs LLM Safety
University of Cagliari · University of Genova
CR4T: Rewrite-Based Guardrails for Adolescent LLM Safety
May 20, 2026Heajun An, Qi Zhang, Vedanth Achanta +1LLM Guardrails LLM Safety Alignment
Virginia Tech
Benchmarking and Improving Monitors for Out-Of-Distribution Alignment Failure in LLMs
May 20, 2026Dylan Feng, Pragya Srivastava, Anca Dragan +1Language Model Safety Evaluation LLM Guardrails
University of California, Berkeley, USA · 3Google DeepMind, India.
REFLECTOR: Internalizing Step-wise Reflection against Indirect Jailbreak
May 20, 2026Jiachen Ma, Jiawen Zhang, Xiangtian Li +3LLM Self-Refinement LLM Safety
Fudan University · Shanghai Artificial Intelligence Laboratory · Zhejiang University
Robotics-Inspired Guardrails for Foundation Models in Socially Sensitive Domains
May 19, 2026Rebecca Ramnauth, Drazen Brscic, Brian ScassellatiLLM Guardrails Human-AI Interaction
Yale University, USA · Kyoto University, Japan
Conformal Selective Acting: Anytime-Valid Risk Control for RLVR-Trained LLMs
May 18, 2026Hamed Khosravi, Xiaoming HuoSelective Prediction Anytime-Valid Inference
Georgia Institute of Technology
Safety Geometry Collapse in Multimodal LLMs and Adaptive Drift Correction
May 18, 2026Jiahe Guo, Xiangran Guo, Jiaxuan Chen +6Multimodal Large Language Models LLM Safety
Harbin Institute of Technology · Huawei Technologies Co., Ltd
Babel: Jailbreaking Safety Attention via Obfuscation Distribution Optimized Sampling
May 18, 2026Ziwei Wang, Jing Chen, Ruichao Liang +6LLM Safety LLM Jailbreak Attacks
Wuhan University · Nanyang Technological University · Southeast University +1
DMN: A Compositional Framework for Jailbreaking Multimodal LLMs with Multi-Image Inputs
May 18, 2026Wenzhuo Xu, Zhipeng Wei, Zonghao Ying +4Multimodal Large Language Models LLM Safety
AI Security Lab, Beijing, China · International Computer Science Institute, CA, USA · UC Berkeley, CA, USA +1

May 28, 2026, cs.LGJ/K move · Enter open · S save

Opir: Efficient Multi-Task Safety Classification for Toxicity, Jailbreaks, Hate Speech, and Harmful Content

Ihor Stepanov, Aleksandr Smechov

1Knowledgator, Kyiv, Ukraine · 2Wordcab, New York, USA

Language Model Safety Evaluation LLM Guardrails Toxicity Detection Safety Filtering LLM Safety Jailbreak Detection+1

Open paper AskPDF arXiv

Real-time safety filtering for large language model (LLM) applications requires classifiers that can detect unsafe prompts, toxic language, jailbreak attempts, and unsafe responses without the cost profile of large guardrail models, and that can distinguish benign sensitive text from genuinely covert harmful content. In this paper, we introduce Opir, a family of encoder-based guardrail models built on the GLiClass architecture. Opir includes multi-task models for binary safe/unsafe classification, multi-label toxicity classification, jailbreak classification, and zero-shot unsafe prompt and response categorization. We also release edge variants with fewer than 100M parameters dedicated to binary safe/unsafe categorization. The models are trained on a three-level taxonomy containing 996 categories across 16 top-level labels, 126 mid-level labels, and 854 leaf labels. Opir's training data combines taxonomy-grounded unsafe prompts, adversarially mined hard negatives, benign safety-preserving examples, generated response examples, multilingual translations, and portions of the Aegis2 and WildGuard training subsets. We also open-sourced an evaluation harness that supports GLiClass and GLiNER2 backends as well as decoder-based models, and covers binary safety classification, multi-label categorization, toxicity, jailbreak detection, prompt safety, response safety, response refusal, and prompt subcategory views across public benchmark families. Across an expanded comparison spanning 12 safety-classification tasks and 17 category tasks against eight contemporary guardrail systems -- including both GLiNER2-based and generative guardrail models -- Opir variants are competitive on or ahead of the strongest open-weight baselines on the majority of benchmark datasets while operating with a substantially smaller deployment footprint.

Momentum

Related topics

Latest papers 272