Refusal Behavior in Language Models

Follow this topic Ask about this topic

Momentum

10 papers in the last four weeks, down 33% on the four weeks before. 0.1% of all new papers.

Jul 13Week of Sep 28

Latest papers 93

All topics

When Refusal Looks Safe: The Refusal-Cue Shortcut in Safety Guard Models
Aug 4, 2026Yu Feng, Chunting Zang, Chen Shen +4Language Model Safety Evaluation LLM Guardrails
The University of Sydney · Alibaba Group · Jilin University +1
Aligned in Form, Not in Meaning: The Comprehension - Containment Decoupling of LLM Safety in Low-Resource Bangla Derogatory Speech
Aug 3, 2026Shadab Bin Habib, A K M Ferdous Reza Habib, Subarno Neel +1Language Model Safety Evaluation LLM Safety Alignment
Department of Computer Science and Engineering Islamic University of Technology, Dhaka, Bangladesh
HopRefusalBench: Diagnosing Refusal Failures in Search-Augmented Agents for Multi-Hop Reasoning
Aug 2, 2026Jianan Xie, Xin Sun, Zhongqi Chen +3Multi-Hop QA AI Agent Benchmarks
ShanghaiTech University · NLPR, MAIS, CASIA · Ant Group
Knowing When to Quit: Diagnosing and Training LLMs to Abort Futile Reasoning
Jul 31, 2026Xinyan Guan, Jiali Zeng, Chunlei Xin +5Overthinking in Language Models RL for Language Model Reasoning
1Chinese Information Processing Laboratory, Institute of Software, Chinese Academy of Sciences · University of Chinese Academy of Sciences · 3Weixin AI, Tencent Inc, China
When LLM Defenses Backfire: Characterizing Safety, Performance, and Cost Trade-offs
Jul 27, 2026Tong Zhang, Zexin Li, Simin Chen +1LLM Inference Efficiency Language Model Safety Evaluation
Fudan University · University of California, Riverside · Columbia University
Asymmetric Collapse in Model Merging: When Refusal Over- writes Recognition
Jul 26, 2026Aarnav Choudhary, Matheus Fonseca Rocha, Jiwon Seo +2Language Model Safety Evaluation Task Vector Merging
UCLA · Algoverse · Purdue University +1
Do LLMs Know Their Vulnerable Scenarios?
Jul 26, 2026Ziheng Peng, Huiqi Deng, Haoran Jing +5LLM Interpretability Adversarial Scenario Generation
Xi’an Jiaotong University Shanghai, China · Shanghai Artificial Intelligence Laboratory Shanghai, China
Refusal-Gated Decoding: Preserving Refusal Behavior Under High-Temperature Sampling
Jul 22, 2026Phillip Howard, Xin Su, Allen Roush +3Language Model Decoding LLM Refusal Behavior
Thoughtworks
Mark, Don't Erase: Token Inoculation for Dual-Use Knowledge in LLMs
Jul 21, 2026Seunghyun Lee, Dongyoon Han, Sangdoo YunLLM Alignment LLM Refusal
NAVER AI Lab
Abliteration Is Not a Scalpel: Off-Target Effects of Refusal Removal on Decision Disposition Across Model Families
Jul 19, 2026Aleksander FafułaLLM Auditing LLM Refusal Behavior
independent researcher • funduszai.pl
Capabilities of Claude Fable 5 on Biomedical Challenge Problems
Jul 12, 2026Dominic Okonkwo, Magnus Hodgson, Temitope I. David +1Biomedical QA LLM Refusal Behavior
School of Computing, University of Georgia · Department of Chemistry, University of Illinois · Institute of Bioinformatics, University of Georgia
Two Axes of LLM Abstention: Answer Correctness and Question Answerability
Jul 9, 2026Benedikt J. WagnerSelective Prediction Confidence Estimation in Language Models
City St George’s, University of London
Faithfulness to Refusal: A Causal Audit of Neuron Selectors
Jul 6, 2026Ananth Eswar, Pratinav Seth, Utsav Avaiya +1Transformer Interpretability Feature Attribution
Lexsi Labs · Work done while at Lexsi Labs.
Fast Multi-dimensional Refusal Subspaces via RFM-AGOP
Jul 2, 2026Thomas WinningerLLM Interpretability LLM Safety
Télécom SudParis, Évry-Courcouronnes, France · ENS Paris-Saclay, Gif-sur-Yvette, France
Addressing Over-Refusal in LLMs with Competing Rewards
Jun 30, 2026Taeyoun Kim, Aviral KumarRL for Language Model Reasoning LLM Safety
Carnegie Mellon University · SynthLabs
RAS: Measuring LLM Safety Through Refusal Alignment
Jun 24, 2026Chang-Chieh Huang, Yan-Lun Chen, Chia-Mu Yu +1LLM Safety Alignment LLM Safety
National Yang Ming Chiao Tung University · Hon Hai Research Institute
Refusal Lives Downstream of Persona in Chat Models
Jun 24, 2026Viola Zhong, Qirui LiLanguage Model Steering LLM Refusal Behavior
Independent · Department of Mathematics Pohang University of Science and Technology
LLMs Prompted for Legal Context Object More: Overrefusal from Small On-Premises LLMs in Criminal Legal Context
Jun 23, 2026Anastasiia Kucherenko, François Brouchoud, Dimitri Percia David +1LLM Prompting Small Language Models
IEM, HEG, HES-SO Valais-Wallis Sierre, Switzerland
Skin-Deep: A Geometric Diagnostic for Alignment Fragility in Large Language Model Representations
Jun 21, 2026Dongyub Jude Lee, Jungseob Lee, Seungyoon Lee +5LLM Interpretability LLM Fine-Tuning
Zoom Communications · Korea University · Yonsei University +1
Safe responses matter: Output-aware safety guardrail mitigate over-refusal in MLLMs
Jun 20, 2026Jiayi Li, Kun ZhanLLM Guardrails Multimodal Large Language Models
Lanzhou University
AOR-Bench: Do Large Audio Language Models Over-Refuse Pseudo-Harmful Queries?
Jun 19, 2026Jiaxi Yang, Chaewan Chun, Jason Lucas +2Audio-Language Model Evaluation LLM Safety
Penn State University
LoMC: Localized Multidirectional Correction for Refusal Suppression in Routed Foundation Models
Jun 10, 2026Yan Hong, Kedong Xiu, Wei Li +6Mixture-of-Experts Language Models LLM Refusal Behavior
Ant Group · Zhejiang University · Shanghai Jiao Tong University
Quality Is Not a Safety Proxy Under Quantization
Jun 8, 2026Sahil KadadekarLLM Quantization LLM Refusal Behavior
Independent Researcher
PsychoSafe: Eliciting Psychologically-Informed Refusals in Large Language Models
Jun 8, 2026Gianluca Barmina, Federico Torrielli, Sven Harms +7Emotional Support Conversation Mental Health
University of Southern Denmark · University of Turin · University of Hamburg +1
Expert-Aware Refusal Steering
Jun 2, 2026Anna C. Marbut, Daniel R. Olson, Travis J. WheelerMixture-of-Experts Language Models Language Model Steering
Department of Interdisciplinary Studies University of Montana Missoula, MT, USA · Department of Pharmacy Practice & Science University of Arizona Tucson, AZ, USA · European Bioinformatics Institute European Molecular Biology Laboratory Wellcome Genome Campus, Hinxton, United Kingdom
DDOR: Delta Debugging for Explainable Overrefusal Testing and Repair
Jun 2, 2026Qinyan Zhou, Peixin Zhang, Jun Sun +2LLM Guardrails LLM Refusal Behavior
Southeast University, China · School of Computing and Information Systems, Singapore Management University, Singapore · Zhejiang University, China +1
TukaBench: A Culturally Grounded Jailbreak Benchmark for African Languages
May 31, 2026Victor Akinode, Senyu Li, Wassim Hamidouche +3LLM Safety Benchmarks Multilingual Language Model Evaluation
Mila - Quebec AI Institute · McGill University · Microsoft AI for Good Research Lab +1
Low-Resource Safety Failures Are Action Failures, Not Representation Failures
May 31, 2026Rashad Aziz, Ikhlasul Akmal Hanif, Fajri KotoLanguage Model Calibration LLM Safety
Mohamed bin Zayed University of Artificial Intelligence
BioRefusalAudit: Auditing Biosecurity Refusal Depth Using General and Domain-Fine-Tuned Sparse Autoencoders
May 28, 2026Caleb DeLeeuwLanguage Model Safety Evaluation Sparse Autoencoders
Independent researcher
Beyond Attack Success Rate: Temporal Logit Observability for LLM Safety Failures
May 28, 2026Junyoung Park, Sunghwan Park, Seongyong Ju +1LLM Refusal Behavior LLM Jailbreak Attacks
Chung-Ang University

Aug 4, 2026, cs.AIJ/K move · Enter open · S save

When Refusal Looks Safe: The Refusal-Cue Shortcut in Safety Guard Models

Yu Feng, Chunting Zang, Chen Shen, Rui Miao+3

The University of Sydney · Alibaba Group · Jilin University+1

Language Model Safety Evaluation LLM Guardrails Safety Filtering Causal Interventions in Language Models Refusal Behavior in Language Models

Open paper AskPDF arXiv

Safety guards are widely used to filter harmful content and are typically trained via supervised fine-tuning on labeled prompt-response pairs. We audit two widely used safety-guard training datasets, WildGuardMix and GR-Train, and find that among responses to harmful prompts, refusal expressions co-occur almost exclusively with unharmful labels. This imbalance motivates what we term the refusal-cue shortcut: inserting a refusal cue into a harmful response could flip the guard's verdict from harmful to unharmful. The shortcut affects not only guards trained on these datasets but also officially released models such as LlamaGuard3 and Qwen3Guard whose training data is undisclosed. It persists across response positions and is generally stronger in smaller variants within a family. To mitigate it, we adapt sparse complementary masking as a lightweight post-hoc intervention that identifies and suppresses a small set of shortcut-associated attention heads and MLP neurons without retraining. On two primary benchmarks, the intervention achieves an approximately 79% relative reduction in response-initial detection failures induced by refusal cues, while preserving standard detection performance. Although optimized using cues at a single response position, the suppression effect transfers to unseen positions and datasets, suggesting that shortcut manifestations across positions are partly mediated by shared internal components. Further analysis provides evidence that shortcut reliance and legitimate refusal recognition are partially functionally separable, as suppressing the shortcut broadly preserves the guard's ability to recognize genuine refusals.

Momentum

Related topics

Latest papers 93