Language Model Safety Evaluation

Follow this topic Ask about this topic

Language model safety evaluation asks whether models handle harmful, ambiguous, and high-stakes requests safely while still helping with legitimate ones. Researchers use benchmarks such as HarmBench, JailbreakBench, and XSTest, alongside static and adaptive red-teaming, expert review, and automated judges; newer tests also examine multi-turn, tool-using, speech, and multimodal systems. Psychometric methods, repeated sampling, and internal activation probes can reveal risks that single-run refusal scores miss, while localized and domain-specific evaluations test language, culture, and settings such as healthcare or finance. A central challenge is making results reliable and predictive of deployment: robustness can shift with wording, language, interaction length, or evaluator, and reducing harmful compliance can increase over-refusal or undermine useful performance.

Latest papers

550 papers82 papers in the last four weeks, up 64% on the four weeks before. 0.5% of all new papers.

How I Aligned Your Model: Scaling Up Multilingual Safety Training with Templates
Oct 29, 2026Max Neuwinger, Hanna Yukhymenko, Imanol SchlagLLM Safety Alignment Language Model Safety Evaluation
ETH Zurich
Red-Teaming Large Language Models in Low-Resource African Languages: A Case Study of Yoruba and Nigerian Pidgin
Oct 29, 2026Paula Chikaodinaka Anyanwu, Valdini Douglace Lemofouet, Blessing Ngozi Uzor +7Language Model Safety Evaluation LLM Red Teaming
African Institute for Mathematical Sciences (AIMS), Cameroon · Independent Researcher · Brown University +5
HEARSAYBENCH: Can AI Navigate from Abstract Human Rights to Lived Lives?
Oct 28, 2026Ava Iranmanesh, Sobhan Lotfi, Ali Iranmanesh +1LLM Safety Benchmarks Language Model Safety Evaluation
Penn State University · Independent Researcher · University of Washington
Rethinking LLM Jailbreak Evaluation: Continuous Risk Landscapes Beyond Binary Outcomes
Oct 28, 2026Yukai Zhou, Wenjie WangLanguage Model Safety Evaluation Jailbreak Attack Evaluation
ShanghaiTech University
When Support Crosses the Line: Evaluating Ethical Boundaries in LLM-Based Mental Health Support in the Chinese Context
Oct 28, 2026Bichen Wang, Yufeng Jin, Yixin Sun +2Mental Health Conversational Agents LLM Safety Benchmarks
Research Center for Social Computing and Information Retrieval, Harbin Institute of Technology, Heilongjiang, China
Masked Erosion: How Acceptance Gates Lose Redundant Safeguards in Self-Improving Agents
Oct 28, 2026Hau-Sing So, Daixian Liu, Weicai LongLLM Agent Safety LLM Agent Self-Improvement
University of Macau · Tsinghua University
Component Translation and Judge-Labelled Refusal: A Multilingual Safety Evaluation
Oct 28, 2026Yongxi Zhou, Junwei Yao, Wenbo Ye +1Language Model Safety Evaluation LLM Safety Evaluation
Independent Researcher
Aligning Large Language Models for Safety in African Languages: A Case Study of Hausa
Oct 27, 2026Valdini Douglace Lemofouet, Blessing Ngozi Uzor, Paula Chikaodinaka Anyanwu +4LLM Safety Alignment Language Model Safety Evaluation
African Institute for Mathematical Sciences (AIMS), Cameroon · Bayero University Kano · University of Hamburg +1
Linear Probes Transfer from Base to Distilled Language Models despite Rotation
Oct 26, 2026Marios TsatsosLanguage Model Probing LLM Safety Monitoring
Independent researcher
Amplify, Locate, Estimate (ALE) on Rare-Failure Rates of Fine-Tuned LLMs
Oct 26, 2026Patrick Reichherzer, Jonathan Evertz, Thorsten HolzRare-Event Estimation Language Model Safety Evaluation
Max Planck Institute for Security and Privacy (MPI-SP) · CISPA Helmholtz Center for Information Security
TrustJev: Separating Policy Authority from Untrusted Evidence in LLM-Based Decisions
Oct 26, 2026Zhiwei Xu, Xuyang Liu, Xiaolin Peng +8LLM Decision-Making LLM Security
KLISS, BNRist, School of Software, Tsinghua University, Beijing, China · ChinaBond Finance and Information Technology Co., Ltd., Beijing, China
Task Success Is Not Enough: Side-effect-Aware Evaluation of Tool-Using Language Model Agents
Oct 25, 2026Jiaju Huang, Shaobin Chen, Xinglong Liang +5LLM Agent Evaluation LLM Agent Safety
Macao Polytechnic University · Southwest Medical University · Netherlands Cancer Institute +1
Identically Trained Language Models Break on Different Jailbreaks
Oct 25, 2026Chi-Wei Chang, Mu-Tien Kuo, Kuan-Wei Lee +1Language Model Safety Evaluation LLM Jailbreak Attacks
Cerulean Labs · National Taiwan University, Taipei, Taiwan · National Yang Ming Chiao Tung University, Hsinchu, Taiwan
Beyond Belief: Evaluating and Mitigating LLM Susceptibility to Misinformation
Oct 25, 2026Emran Kanaan, Jiho Lee, Ibrahim Odat +2Language Model Safety Evaluation Language Model Robustness
Oakland University · Dartmouth College
Task Format as a Factor in Adaptation: A Controlled Comparison of LoRA and Full Fine-Tuning for Prompt-Safety Classification
Oct 24, 2026John Alderete, Srirakshith Betageri, Jennifer Flynn +3Fine-Tuning Language Models Low-Rank Adaptation
Fiddler AI · Simon Fraser University · Georgia Institute of Technology
Propensities Please: Exposing Alignment-Faking LLMs using Synthetic Document Finetuning
Oct 23, 2026Ardy Haroen, Florian Lorkowski, Samuel Nellessen +1Language Model Safety Evaluation LLM Auditing
LASR Labs · Apollo Research
$M^3$ : A Multi-LLM System for Multi-Turn Jailbreaking as a Markov Decision Process
Oct 22, 2026Yuqing Huang, Jack Ji, Jithin Pradeep +2Adversarial Attacks on LLMs Multi-Turn Jailbreak Attacks
The Vanguard Group∗
Simulated Control Evals
Oct 22, 2026Jeffrey Olmo, David Zhang, Tyler TracyAI Safety Evaluation Language Model Safety Evaluation
Redwood Research
How Firm Are AI Models' Ethical Boundaries? A Pay-Stub Editing Test of 15 Models
Oct 22, 2026Simiao Ren, Xingyu Shen, Tommy Duong +2Language Model Safety Evaluation Refusal Behavior in Language Models
Faithful Latent Harm Monitors Work No Better Than the Models Own Behaviour at a Matched False Alarm Rate
Oct 22, 2026Abhishek Mishra, Sagnik Chatterjee, Suyash ManiyarLLM Safety Monitoring Language Model Safety Evaluation
University of Massachusetts Amherst
Beyond End-to-End Recovery: Towards a Better Understanding of LLM Steganography via Rolewise Decomposition
Oct 21, 2026Om Tailor, Yucheng Pan, Solaleh Mohammadi +2Language Model Steganography LLM Security
University of Maryland, College Park
Unintended Side Effects of Value Training in LLMs
Oct 21, 2026Siddhesh Pawar, Chen Cecilia Liu, Anna Korhonen +1LLM Alignment Language Model Post-Training
University of Copenhagen · University of Cambridge
TenetBench: A Clause-Anchored Russian Corpus for Evaluating Language Models against Professional Codes of Ethics
Oct 21, 2026Yegor Anichkov, Valery Vinogradov, Andrey Furov +2Language Model Safety Evaluation LLM Safety Benchmarks
Trusted AI Research Center, RAS, Moscow, Russia. · Center for Applied Artificial Intelligence, Institute of Social Sciences, Moscow, Russia.
Fooling LLM Safety Guards into Judging the Wrong Text
Oct 20, 2026Utkarsh KanwatLLM Guardrails Adversarial Attacks on LLMs
Independent Researcher
Polarity Blindness in LLM Safety: A Systematic Study of Negation, Antonym Substitution, and Adversarial Framing
Oct 20, 2026Diego Miguel Lozano, Erik Derner, Juan Antonio Pérez-Ortiz +1Language Model Safety Evaluation Adversarial Robustness of Language Models
ELLIS Alicante, Spain · Czech Technical University in Prague, Czech Republic · Universidad de Alicante, Spain
ABench: Adversarily Benchmarking Large Language Models across Taxonomized Values
Oct 20, 2026Liping Yang, Yongqi Leng, Jingting Zheng +3LLM Safety Benchmarks Language Model Safety Evaluation
TJUNLP Lab, School of Computer Science and Technology, Tianjin University, Tianjin, China · Yuanhui AI
When LLMs Manipulate: Measuring Behavioral, Social, and Temporal Effects
Oct 20, 2026Sadia Asif, Israel Antonio Rosales Laguan, Shumaila Asif +3Language Model Safety Evaluation Persuasion in Language Models
Rensselaer Polytechnic Institute, USA · Independent Researcher, Colombia · National University of Sciences and Technology, Pakistan +1
Decision Sufficiency of Learned Representation Interfaces for Safe Generative Control
Oct 20, 2026Jipeng Li, Yanning ShenLanguage Model-Based Control Language Model Safety Evaluation
University of California, Irvine
Global Security Blind Spots: Translation as a Safety Bypass in Low-Resource Languages
Oct 19, 2026Jalil Alva, David Harper, Abigail E. Thornton +8LLM Safety Language Model Safety Evaluation
Welo Data
African Languages AI Safety Readiness Report
Oct 19, 2026Israel Abebe Azime, Jesujoba O. Alabi, Tajuddeen Gwadabe +1Language Model Safety Evaluation African Language Processing
Masakhane · Saarland University · Mila - Quebec AI Institute, McGill University, and Canada CIFAR AI Chair

Oct 29, 2026J/K move · Enter open · S save

How I Aligned Your Model: Scaling Up Multilingual Safety Training with Templates

Max Neuwinger, Hanna Yukhymenko, Imanol Schlag

ETH Zurich

LLM Safety Alignment Language Model Safety Evaluation Data Augmentation LLM Safety Evaluation

Open paper AskPDF arXiv

Safety alignment for Large Language Models is heavily focused on English with limited multilingual representation, particularly damaging low-resource language performance. Related open training datasets are often limited in size or do not follow region-localized harmful topics to broaden the scope of alignment beyond cultures. We propose a scalable template-based augmentation pipeline to produce massive multilingual refusal and benign corpora, extracting safe templates from harmful prompts, translating them and then filling in the slots with culturally authentic entities, both harmful and benign. As a result, we release MSafeTemp, a validated corpus spanning 60 languages useful for multilingual safety alignment and evaluation. Using the dataset, we show that English-only evaluation misjudges multilingual safety: for several models, attacks that fail in English succeed once translated, while culturally localized variants expose further failures that translation alone misses. We also confirm that augmenting safety training data through templates allows for higher diversity and quality of alignment, resulting in improved safety performance beyond English. MSafeTemp and the respective pipeline are aimed at facilitating research and development of safer and more responsible models for less-represented cultures.