LLM Safety Evaluation

LLM: Large Language Model

The field tests whether language models refuse harmful requests, handle benign sensitive tasks helpfully, and remain safe when prompts, context, or tools change. Common approaches include fixed safety benchmarks, adversarial red-teaming and jailbreak testing, multi-turn agent simulations, domain-specific scenarios, and automated or human judging. Researchers increasingly measure more than refusal rates or attack success, including harm severity, omissions, false refusals, reasoning and tool-use failures, and effects on users. A central challenge is that results can shift with benchmark design, evaluation conditions, and judge choice, while stronger safeguards may reduce helpfulness and static tests may not predict behavior in deployment.

Latest papers

261 papers29 papers in the last four weeks, up 12% on the four weeks before. 0.2% of all new papers.

A Systematic Study of Training-Free Methods for Trustworthy Large Language Models
Apr 17, 2026Wai Man Si, Mingjie Li, Michael Backes +1LLM Alignment LLM Safety
CISPA Helmholtz Center for Information Security
A Multi-Domain Red Teaming Framework for Safety, Robustness, and Fairness Evaluation of Medical Large Language Models
Apr 15, 2026Andrei Marian Feier, Veysel Kocaman, Yigit Gul +6LLM Safety LLM Safety Evaluation
John Snow Labs Inc.
Calibrate-Then-Delegate: Safety Monitoring with Risk and Budget Guarantees via Model Cascades
Apr 15, 2026Edoardo Pona, Milad Kazemi, Mehran Hosseini +4LLM Safety Evaluation
King’s College London · University of Manchester · Northeastern University London
Model Capability Assessment and Safeguards for Biological Weaponization
Apr 15, 2026Michael RichterResponsible AI LLM Safety Evaluation
Binghamton University
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
ContextLens: Modeling Imperfect Privacy and Safety Context for Legal Compliance
Apr 14, 2026Haoran Li, Yulin Chen, Huihao Jing +6LLM Safety Evaluation AI Assurance
Beihang University · HKUST · National University of Singapore +1
Do LLMs Follow Their Own Rules? A Reflexive Audit of Self-Stated Safety Policies
Apr 10, 2026Avni MittalLLM Auditing LLM Safety Evaluation
Evaluating Large Language Models in a Complex Hidden Role Game
Apr 9, 2026Niklas BauerDeception in Language Models Strategic Reasoning Risks in Language Models
University of Göttingen
IatroBench: A Pre-Registered Benchmark of Clinical Omission in Language Models
Apr 9, 2026David GringrasLLM Safety Benchmarks Healthcare
Harvard T.H. Chan School of Public Health
LURE: Live-Usage Replay Evaluations for Reducing Evaluation Awareness
Apr 8, 2026Igor Ivanov, David Demitri AfricaLLM Evaluation LLM Safety Evaluation
Meridian Cambridge
GrandGuard: Taxonomy, Benchmark, and Safeguards for Elderly-Chatbot Interaction Safety
Apr 7, 2026Changxuan Fan, Xi Yang, Yueyuan Zheng +9LLM Safety Evaluation LLM Safety Benchmarks
The Hong Kong University of Science and Technology
Beyond Idealized Patients: Evaluating LLMs under Challenging Patient Behaviors in Medical Consultations
Mar 31, 2026Yahan Li, Xinyi Jie, Wanjia Ruan +5Healthcare Clinical Reasoning
University of Southern California
Peer-Preservation in Frontier Models
Mar 30, 2026Yujin Potter, Nicholas Crispino, Vincent Siu +2AI Agent Safety LLM Safety Evaluation
University of California, Berkeley · University of California, Santa Cruz
The Geometry of Harmful Intent: Training-Free Anomaly Detection via Angular Deviation in LLM Residual Streams
Mar 28, 2026Isaac Llorente-SaguerAnomaly Detection Language Model Safety Evaluation
Independent Researcher
LLM Readiness Harness: Evaluation, Observability, and CI Gates for LLM/RAG Applications
Mar 28, 2026Alexandre Cristovão MaioranoLLM Evaluation RAG Evaluation
Lumytics
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
LJ-Bench: Ontology-Based Benchmark for U.S. Crime
Mar 21, 2026Hung Yun Tseng, Wuzhen Li, Blerina Gkotse +1LLM Safety Benchmarks LLM Safety Evaluation
University of Wisconsin–Madison
Using LLM-as-a-Judge/Jury to Advance Scalable, Clinically-Validated Safety Evaluations of Model Responses to Users Demonstrating Psychosis
Mar 20, 2026May Lynn Reese, Markela Zeneli, Mindy Ng +3LLM Safety LLM Safety Evaluation
Apart Research · Odyssean Institute · London School of Economics and Political Science +1
Toward Reliable, Safe, and Secure LLMs for Scientific Applications
Mar 18, 2026Saket Sanjeev Chaturvedi, Joshua Bergerson, Tanwi MallickLLM Safety AI Agent Safety
Argonne National Laboratory, Lemont, 60439, USA
Multi-Trait Subspace Steering to Reveal the Dark Side of Human-AI Interaction
Mar 18, 2026Xin Wei Chia, Swee Liang Wong, Jonathan PanHuman-AI Interaction LLM Safety
Home Team Science & Technology Agency (HTX) Singapore
SafeTutors: Benchmarking Pedagogical Safety in AI Tutoring Systems
Mar 18, 2026Rima Hazra, Bikram Ghuku, Ilona Marchenko +5Intelligent Tutoring Systems AI Safety Evaluation
Eindhoven University of Technology, Netherlands (TU/e) · Indian Institute of Technology Kharagpur · Taras Shevchenko National University of Kyiv +3
Evolving Contextual Safety in Multi-Modal Large Language Models via Inference-Time Self-Reflective Memory
Mar 16, 2026Ce Zhang, Jinxi He, Junyi He +2Multimodal Large Language Models Language Model Safety Evaluation
Robotics Institute, Carnegie Mellon University
Beyond Creed: A Non-Identity Safety Condition A Strong Empirical Alternative to Identity Framing in Low-Data LoRA Fine-Tuning
Mar 16, 2026Xinran ZhangSafety-Preserving Fine-Tuning of Language Models LLM Safety
University of California, Berkeley
Exposing Long-Tail Safety Failures in Large Language Models through Efficient Diverse Response Sampling
Mar 15, 2026Suvadeep Hajra, Palash Nandi, Tanmoy ChakrabortyLLM Safety LLM Red Teaming
Department of Electrical Engineering, Indian Institute of Technology Delhi · Yardi School of Artificial Intelligence, Indian Institute of Technology Delhi
QuarkMedBench: A Real-World Scenario Driven Benchmark for Evaluating Large Language Models
Mar 14, 2026Yao Wu, Kangping Yin, Liang Dong +13LLM Evaluation Clinical Language Model Evaluation
Quark Medical Team, Alibaba Group
Risk-Adjusted Harm Scoring for Automated Red Teaming for LLMs in Financial Services
Mar 11, 2026Fabrizio Dimino, Bhaskarjit Sarmah, Stefano PasqualiLanguage Model Safety Evaluation LLM Safety
1Domyn, New York, US · 2Domyn, New York, US · 3Domyn, New York, US
Disentangling Prompt Element Level Risk Factors for Hallucinations and Omissions in Mental Health LLM Responses
Mar 10, 2026Congning Ni, Sarvech Qadir, Bryan Steitz +14Hallucination in Language Models LLM Safety Evaluation
Vanderbilt University Medical Center, Nashville, TN, USA · Vanderbilt University, Nashville, TN, USA · Lirio, Knoxville, TN, USA +1
Safety Under Scaffolding: How Evaluation Conditions Shape Measured Safety
Mar 8, 2026David GringrasLLM Safety Benchmarks LLM Agent Evaluation
Harvard University
Evaluating Human-AI Safety: A Framework for Measuring Harmful Capability Uplift
Mar 6, 2026Michelle Vaccaro, Jaeyoon Song, Abdullah Almaatouq +1AI Safety Evaluation LLM Safety Evaluation
Massachusetts Institute of Technology, Cambridge, USA
GMP: A Benchmark for Content Moderation under Co-occurring Violations and Dynamic Rules
Mar 2, 2026Houde Dong, Yifei She, Kai Ye +3Content Moderation LLM Safety Evaluation
Beijing University of Posts and Telecommunications · The University of Hong Kong

Apr 17, 2026J/K move · Enter open · S save

A Systematic Study of Training-Free Methods for Trustworthy Large Language Models

Wai Man Si, Mingjie Li, Michael Backes, Yang Zhang

CISPA Helmholtz Center for Information Security

LLM Alignment LLM Safety LLM Safety Evaluation Language Model Robustness

Open paper AskPDF arXiv

As Large Language Models (LLMs) receive increasing attention and are being deployed across various domains, their potential risks, including generating harmful or biased content, producing unsupported claims, and exhibiting vulnerabilities to adversarial attacks, have drawn significant attention. To enable quick and low-cost adaptation, training-free methods have recently emerged as cost-effective alternatives to post-training alignment techniques. Despite their promising results, these methods are evaluated inconsistently across the literature, cover limited dimensions of trustworthiness, and can introduce undesirable side effects, such as utility degradation and increased brittleness. To fully assess the impacts of these training-free methods, we take a step back and systematically re-evaluate the effectiveness of existing training-free methods against various trustworthy settings and their influence on utility, robustness, and computational overhead. We also categorize these methods into three levels (input, internal, and output) based on where they intervene in the model's information flow during inference. Using this taxonomy, we conduct a comprehensive analysis of various representative and effective methods from each level across different LLM families and sizes. Our analysis highlights several trade-offs and unresolved challenges in current approaches. We summarize key findings and limitations in the existing literature, and propose practical recommendations for balancing trustworthiness, utility, and robustness in LLMs without the need for additional training.