Language Model Safety Evaluation

Follow this topic Ask about this topic

Momentum

45 papers in the last four weeks, down 15% on the four weeks before. 0.3% of all new papers.

Jul 13Week of Sep 28

Latest papers 396

All topics

A Translational Note on AI Safety Evaluation
Sep 6, 2026Madhava GaikwadLanguage Model Safety Evaluation AI Safety Evaluation
Independent Researcher, Bengaluru, India
SRD-GUARD: A Defense Framework of LLMs via Semantic Rewriting and Joint Multi-Model Scoring for Latent Intent Exposure
Sep 6, 2026Qi Wang, Chengcheng Wan, Jiangtao WangLLM Guardrails LLM Jailbreak Attacks
East China Normal University Shanghai, China · East China Normal University, Shanghai Innovation Institute Shanghai, China · Software Engineering Institute, East China Normal University Shanghai, China
Beyond the Flag: Clinical Framing Closes the Moderation Gap in Suicide Risk Measurement
Sep 5, 2026Shreyas Krishnan, Gun Ahn, Jungjin KimSuicide Risk Modeling Content Moderation
University of California, Berkeley · Wondi AI · MIT +2
SAFEGuard: Detect Optimization-Based Jailbreak Attacks Through Harmful Semantic Analysis and Fluency Measurement
Sep 5, 2026Quoc Viet Vo, Trung Le, Damith C. Ranasinghe +1LLM Security Jailbreak Attacks
Australian Institute for Machine Learning · Adelaide University · Monash University
SinoGlyphBench: A Diagnostic Benchmark for Chinese Glyph-Level Obfuscation in Language-Model Moderation
Sep 5, 2026Yifan Wang, Zimu Wang, Suliu Qin +10Language Model Safety Evaluation Content Moderation
East China Normal University · New York University Shanghai · Xi’an Jiaotong-Liverpool University +3
Safety Monitors Mostly Catch What the Model Already Refuses
Sep 5, 2026Sripad KarneLanguage Model Safety Evaluation LLM Guardrails
Columbia University
Language models judge war differently when tested for alignment
Sep 4, 2026Maxim ChupilkinEvaluation Awareness in Language Models Language Model Safety Evaluation
Department of Politics and International Relations, University of Oxford
Refuse without Refusal: A Structural Analysis of Safety-Tuning Responses for Reducing False Refusals in Language Models
Sep 4, 2026Minji Kim, Hyounghun KimRefusal Behavior in Language Models Fine-Tuning
Graduate School of Artificial Intelligence, POSTECH · Department of Computer Science and Engineering, POSTECH
Safety for Whom? Boundary-Aware Self-Distillation for Controlled LLM Safety Refusal
Sep 3, 2026Alejo López-Ávila, Iker García-Ferrero, Jezabel Garcia +2LLM Safety Alignment Refusal Behavior in Language Models
Multiverse Computing
Probing Warmth-Mediated Harm in Speech-Enabled LLMs for Mental-Health Conversations
Sep 1, 2026Eugenia Kim, Bolor-Erdene Jagdagdorj, Dina Pekelis +2Audio-Language Model Evaluation Language Model Safety Evaluation
Microsoft, Redmond, WA, USA
Who Judges the Judges? A Chinese Safety QA Benchmark for Evaluating LLM Responses and Safety Judges
Sep 1, 2026Rui Yang, Shuang Huang, Junhua Liu +7LLM Safety Benchmarks LLM-as-a-Judge
Anhui SparkShield Intelligent Technology · iFLYTEK · Peking University +1
HiveTraceGuard-Pro: A Compact Generative Guardrail for Prompt Injection, Jailbreaks, and Adversarial Obfuscation
Sep 1, 2026Nikita Oblakov, Sabrina Sadiekh, Evgeniy KokuykinLanguage Model Safety Evaluation LLM Guardrails
HiveTraceLab
Validity-Aware Jailbreak Evaluation for Large Language Models
Aug 31, 2026Qilong Wu, Sahil Wadhwa, Pranab Mohanty +2Language Model Safety Evaluation LLM Auditing
University of Illinois Urbana-Champaign · Capital One
Beyond Token Positions: Safety Alignment Across Denoising Steps in Diffusion Language Models
Aug 31, 2026Guoli Wang, Haonan Shi, Tu Ouyang +1Language Model Safety Evaluation Diffusion Language Model Decoding
Case Western Reserve University 10900 Euclid Avenue, Cleveland, OH 44106, USA
EvoFlint: An Evolutionary Atlas of Multi-Turn LLM Vulnerabilities
Aug 31, 2026Feitong Qiao, Liren Peng, Shiming Ren +7Language Model Safety Evaluation LLM Red Teaming
Reinforce Labs USA
Capability-Gated Language Models: Security Composes, Utility Does Not
Aug 31, 2026Patrikas Vanagas, Augustas Mačijauskas, Laurynas LopataLanguage Model Safety Evaluation LLM Security
BPTI Vilnius, Lithuania · Independent Researcher Vilnius, Lithuania · askEarth AG Zürich, Switzerland
BLOOM-WILT: Logit Tilting for Behaviour Elicitation in Automated LLM Auditing
Aug 31, 2026Adrians Skapars, Edoardo ManinoLanguage Model Steering Language Model Safety Evaluation
University of Manchester, United Kingdom
You Shouldn't Have Asked: A Pragmatics-Inspired Taxonomy for Evaluating LLM Refusals
Aug 31, 2026Ruoxuan Li, Pinqiao Wang, Sheng Li +1Language Model Safety Evaluation Pragmatic Reasoning in Language Models
Stony Brook University · University of Virginia
The Fragility of Jailbreak Robustness Across Operational States
Aug 31, 2026Yuna Park, Hwang Youn Kim, Yujin Kim +3Prompt Sensitivity Language Model Safety Evaluation
Yonsei University · Korea Institute of Science and Technology (KIST) · Korea University +1
SingProbe Technical Report
Aug 31, 2026Sing TeamLanguage Model Safety Evaluation LLM Guardrails
AI Security Lab, Ant Group
WildSEEK: Evaluating Language Models for Information-Seeking
Aug 31, 2026Tanise Ceron, Joachim Baumann, Elisa Bassignana +3LLM Safety Benchmarks LLM Sycophancy
1Bocconi University · 2Stanford University · 3IT University of Copenhagen +1
Reference-Grafting Matches Fine-Tuning at Eliciting Sandbagged Capabilities
Aug 29, 2026Linh Le, Hong Kiat Tan, David Williams-KingCausal Interventions in Language Models Language Model Safety Evaluation
Lida Safety · University of California, Los Angeles · ERA
Arabic Safety Alignment as Selective Refusal: An Empirical Study of SFT, DPO, and Guard Calibration
Aug 29, 2026Mohamad Zbib, Ammar MohannaRefusal Behavior in Language Models LLM Alignment
American University of Beirut
Not the Same Protector: Deployment-Dependent Protective Intervention in LLMs
Aug 29, 2026Eunna Lee, Soomyoung Lee, Jungpyo Nam +6LLM Safety Language Model Safety Evaluation
Independent Researcher · AI Systems & Data Researcher Art and ArtTech Lee AI Lab · Machine Learning Researcher Recommender Systems Researcher Gachon University +4
REINS: Refusal-Enhanced Inhibitory Steering with Sparse Autoencoder Features
Aug 28, 2026Kai-Xuan Ding, Hao-Xiang Xu, Ji-Hua Peng +3LLM Safety Refusal Behavior in Language Models
University of Science and Technology of China · Zhejiang University
EvoHarmBench: Breaking Content Moderation with Iterative Human-Like Evasion
Aug 28, 2026Ruijie Jian, Benlei Cui, Ting Ma +8Language Model Safety Evaluation LLM Safety Benchmarks
Yuvion Team, Alibaba Group · University of Chinese Academy of Sciences
LongGuard: Mechanistic Analysis and Training-Free Mitigation of Long-Context Failure in Safety Guardrails
Aug 27, 2026Ziyang Chen, Xing Wu, Songlin HuLanguage Model Safety Evaluation Long-Context Language Model Evaluation
Institute of Information Engineering, Chinese Academy of Sciences, Beijing, China · School of Cyber Security, University of Chinese Academy of Sciences, Beijing, China
Nemotron 3.5 Content Safety Moderator: A Compact Multimodal, Multilingual, and Reasoning Enabled Content Safety Moderator
Aug 27, 2026Varun Singh, Anuj Doshi, Makesh Narsimhan Sreedhar +2LLM Guardrails LLM Safety
NVIDIA Santa Clara, CA
NeuronFuzz: Safety Neuron Guided Fuzzing for LLM Safety Evaluation
Aug 26, 2026Zhiyuan Xu, Muhammad Firhard Roslan, Joseph Gardiner +2Language Model Safety Evaluation LLM Security
University of Bristol
SafeTune: A Unified Faithful Library for Auditing and Repairing Safety Drift in Fine-Tuned LLMs
Aug 26, 2026Pratinav Seth, Saisab Sadhu, Anshul Kaushal +1LLM Safety LLM Auditing
Lexsi Labs

Sep 6, 2026, cs.AIJ/K move · Enter open · S save

A Translational Note on AI Safety Evaluation

Madhava Gaikwad

Independent Researcher, Bengaluru, India

Language Model Safety Evaluation AI Safety Evaluation LLM Safety Benchmarks Human-in-the-Loop Evaluation

Open paper AskPDF arXiv

Recent studies report that automated red-teaming finds more vulnerabilities, at lower cost, than human red-teaming on standard AI safety benchmarks, and some read this as evidence that human evaluators are becoming dispensable. The comparison measures one thing and the conclusion claims another. A benchmark measures how thoroughly an attacker searches a predefined set of harms, fixed in advance by the developers, and a harm left out of that set is invisible to any attacker working inside it, automated or not. The same blind spot appeared in academic cryptography and in clinical drug trials, where an evaluation that was internally valid stayed silent about the population it was never pointed at. We call the AI-safety version the \emph{threat-model coverage gap}, and find that it persists in a current open-weight model, where harms surface in non-English prompts that English benchmarks miss. Closing it requires evaluators whose deployment context differs from the developers'. The case for those evaluators is methodological, grounded in coverage, and the existing evaluation frame is unlikely to produce them on its own.

Momentum

Related topics

Latest papers 396