Image de couverture de livre

Universal Refusal Circuits Across LLMs: Cross-Model Transfer via Trajectory Replay and Concept-Basis Reconstruction

Fuente: arXiv

Enregistré dans:

Détails bibliographiques
Auteur principal:	Cristofano, Tony
Format:	Preprint
Publié:	2026
Sujets:	Computation and Language
Accès en ligne:	Acceder al recurso
Tags:	Ajouter un tag Pas de tags, Soyez le premier à ajouter un tag!

Documents similaires

Surgical Refusal Ablation: Disentangling Safety from Intelligence via Concept-Guided Spectral Cleaning
par: Cristofano, Tony
Publié: (2026)

Refusal Direction is Universal Across Safety-Aligned Languages
par: Wang, Xinpeng, et autres
Publié: (2025)

Self and Cross-Model Distillation for LLMs: Effective Methods for Refusal Pattern Alignment
par: Li, Jie, et autres
Publié: (2024)

Differentiable Faithfulness Alignment for Cross-Model Circuit Transfer
par: Shao, Shun, et autres
Publié: (2026)

Refuse Whenever You Feel Unsafe: Improving Safety in LLMs via Decoupled Refusal Training
par: Yuan, Youliang, et autres
Publié: (2024)

LLMs Encode Harmfulness and Refusal Separately
par: Zhao, Jiachen, et autres
Publié: (2025)

RepIt: Steering Language Models with Concept-Specific Refusal Vectors
par: Siu, Vincent, et autres
Publié: (2025)

Refusal in LLMs is an Affine Function
par: Marshall, Thomas, et autres
Publié: (2024)

Over-Refusal and Representation Subspaces: A Mechanistic Analysis of Task-Conditioned Refusal in Aligned LLMs
par: Maskey, Utsav, et autres
Publié: (2026)

Beyond Over-Refusal: Scenario-Based Diagnostics and Post-Hoc Mitigation for Exaggerated Refusals in LLMs
par: Yuan, Shuzhou, et autres
Publié: (2025)

AgentTrek: Agent Trajectory Synthesis via Guiding Replay with Web Tutorials
par: Xu, Yiheng, et autres
Publié: (2024)

Benchmarking Concept-Spilling Across Languages in LLMs
par: Badanin, Ilia, et autres
Publié: (2026)

Think Before Refusal : Triggering Safety Reflection in LLMs to Mitigate False Refusal Behavior
par: Si, Shengyun, et autres
Publié: (2025)

RAID: Refusal-Aware and Integrated Decoding for Jailbreaking LLMs
par: Nguyen, Tuan T., et autres
Publié: (2025)

$C$-$ΔΘ$: Circuit-Restricted Weight Arithmetic for Selective Refusal
par: Kasliwal, Aditya, et autres
Publié: (2026)

Tracing the Dynamics of Refusal: Exploiting Latent Refusal Trajectories for Robust Jailbreak Detection
par: Hu, Xulin, et autres
Publié: (2026)

Can LLMs Refuse Questions They Do Not Know? Measuring Knowledge-Aware Refusal in Factual Tasks
par: Pan, Wenbo, et autres
Publié: (2025)

The Geometry of Refusal in Large Language Models: Concept Cones and Representational Independence
par: Wollschläger, Tom, et autres
Publié: (2025)

Language Model Circuits Are Sparse in the Neuron Basis
par: Arora, Aryaman, et autres
Publié: (2026)

Silenced Biases: The Dark Side LLMs Learned to Refuse
par: Himelstein, Rom, et autres
Publié: (2025)

Learning to Refuse: Towards Mitigating Privacy Risks in LLMs
par: Liu, Zhenhua, et autres
Publié: (2024)

Cannot or Should Not? Automatic Analysis of Refusal Composition in IFT/RLHF Datasets and Refusal Behavior of Black-Box LLMs
par: von Recum, Alexander, et autres
Publié: (2024)

Can Video LLMs Refuse to Answer? Alignment for Answerability in Video Large Language Models
par: Yoon, Eunseop, et autres
Publié: (2025)

Measuring and Enhancing Trustworthiness of LLMs in RAG through Grounded Attributions and Learning to Refuse
par: Song, Maojia, et autres
Publié: (2024)

Refusal Tokens: A Simple Way to Calibrate Refusals in Large Language Models
par: Jain, Neel, et autres
Publié: (2024)

Do Reasoning LLMs Refuse What They Infer in Long Contexts?
par: Fu, Yu, et autres
Publié: (2026)

Cross-model Transferability among Large Language Models on the Platonic Representations of Concepts
par: Huang, Youcheng, et autres
Publié: (2025)

AgentHER: Hindsight Experience Replay for LLM Agent Trajectory Relabeling
par: Ding, Liang
Publié: (2026)

Writer-R1: Enhancing Generative Writing in LLMs via Memory-augmented Replay Policy Optimization
par: Zhao, Jihao, et autres
Publié: (2026)

Understanding Refusal in Language Models with Sparse Autoencoders
par: Yeo, Wei Jie, et autres
Publié: (2025)

Contrastive Cross-Course Knowledge Tracing via Concept Graph Guided Knowledge Transfer
par: Han, Wenkang, et autres
Publié: (2025)

Transferring Expert Cognitive Models to Social Robots via Agentic Concept Bottleneck Models
par: Zhao, Xinyu, et autres
Publié: (2025)

Does Refusal Training in LLMs Generalize to the Past Tense?
par: Andriushchenko, Maksym, et autres
Publié: (2024)

SafeConstellations: Mitigating Over-Refusals in LLMs Through Task-Aware Representation Steering
par: Maskey, Utsav, et autres
Publié: (2025)

RefusalBench: Generative Evaluation of Selective Refusal in Grounded Language Models
par: Muhamed, Aashiq, et autres
Publié: (2025)

Towards Understanding and Improving Refusal in Compressed Models via Mechanistic Interpretability
par: Chhabra, Vishnu Kabir, et autres
Publié: (2025)

Characterizing Selective Refusal Bias in Large Language Models
par: Khorramrouz, Adel, et autres
Publié: (2025)

WildGuard: Open One-Stop Moderation Tools for Safety Risks, Jailbreaks, and Refusals of LLMs
par: Han, Seungju, et autres
Publié: (2024)

Surgical, Cheap, and Flexible: Mitigating False Refusal in Language Models via Single Vector Ablation
par: Wang, Xinpeng, et autres
Publié: (2024)

FalseReject: A Resource for Improving Contextual Safety and Mitigating Over-Refusals in LLMs via Structured Reasoning
par: Zhang, Zhehao, et autres
Publié: (2025)