Image de couverture de livre

Unraveling LLM Jailbreaks Through Safety Knowledge Neurons

Fuente: arXiv

Enregistré dans:

Détails bibliographiques
Auteurs principaux:	Zhao, Chongwen, Ke, Yutong, Huang, Kaizhu
Format:	Preprint
Publié:	2025
Sujets:	Artificial Intelligence
Accès en ligne:	Acceder al recurso
Tags:	Ajouter un tag Pas de tags, Soyez le premier à ajouter un tag!

Documents similaires

Defending against Jailbreak through Early Exit Generation of Large Language Models
par: Zhao, Chongwen, et autres
Publié: (2024)

SafeBehavior: Simulating Human-Like Multistage Reasoning to Mitigate Jailbreak Attacks in Large Language Models
par: Zhao, Qinjian, et autres
Publié: (2025)

Logic Jailbreak: Efficiently Unlocking LLM Safety Restrictions Through Formal Logical Expression
par: Peng, Jingyu, et autres
Publié: (2025)

Efficient LLM-Jailbreaking via Multimodal-LLM Jailbreak
par: Ji, Haoxuan, et autres
Publié: (2024)

Light Alignment Improves LLM Safety via Model Self-Reflection with a Single Neuron
par: Shen, Sicheng, et autres
Publié: (2026)

Toward Principled LLM Safety Testing: Solving the Jailbreak Oracle Problem
par: Lin, Shuyi, et autres
Publié: (2025)

Partial Differential Equations is All You Need for Generating Neural Architectures -- A Theory for Physical Artificial Intelligence Systems
par: Guo, Ping, et autres
Publié: (2021)

Jailbreaking LLM-Controlled Robots
par: Robey, Alexander, et autres
Publié: (2024)

LLM-Powered Explanations: Unraveling Recommendations Through Subgraph Reasoning
par: Shi, Guangsi, et autres
Publié: (2024)

Learning from Risk: LLM-Guided Generation of Safety-Critical Scenarios with Prior Knowledge
par: Wang, Yuhang, et autres
Publié: (2025)

How Alignment and Jailbreak Work: Explain LLM Safety through Intermediate Hidden States
par: Zhou, Zhenhong, et autres
Publié: (2024)

Knowledge-to-Jailbreak: Investigating Knowledge-driven Jailbreaking Attacks for Large Language Models
par: Tu, Shangqing, et autres
Publié: (2024)

Adaptive Prompt Embedding Optimization for LLM Jailbreaking
par: Li, Miles Q., et autres
Publié: (2026)

A generalizable framework for low-rank tensor completion with numerical priors
par: Yuan, Shiran, et autres
Publié: (2023)

Can MLLMs Absorb Math Reasoning Abilities from LLMs as Free Lunch?
par: Hu, Yijie, et autres
Publié: (2025)

Subtoxic Questions: Dive Into Attitude Change of LLM's Response in Jailbreak Attempts
par: Zhang, Tianyu, et autres
Publié: (2024)

Unmasking the Canvas: A Dynamic Benchmark for Image Generation Jailbreaking and LLM Content Safety
par: Nair, Variath Madhupal Gautham, et autres
Publié: (2025)

Align to Misalign: Automatic LLM Jailbreak with Meta-Optimized LLM Judges
par: Koo, Hamin, et autres
Publié: (2025)

Depth Charge: Jailbreak Large Language Models from Deep Safety Attention Heads
par: Wu, Jinman, et autres
Publié: (2026)

MRJ-Agent: An Effective Jailbreak Agent for Multi-Round Dialogue
par: Wang, Fengxiang, et autres
Publié: (2024)

Efficient Safety Retrofitting Against Jailbreaking for LLMs
par: Garcia-Gasulla, Dario, et autres
Publié: (2025)

Jailbreaking Attacks vs. Content Safety Filters: How Far Are We in the LLM Safety Arms Race?
par: Xin, Yuan, et autres
Publié: (2025)

Enhancing Model Defense Against Jailbreaks with Proactive Safety Reasoning
par: Yang, Xianglin, et autres
Publié: (2025)

The Echo Chamber Multi-Turn LLM Jailbreak
par: Alobaid, Ahmad, et autres
Publié: (2026)

Defending Large Language Models Against Jailbreak Attacks via In-Decoding Safety-Awareness Probing
par: Zhao, Yinzhi, et autres
Publié: (2026)

Jailbreaking to Jailbreak
par: Kritz, Jeremy, et autres
Publié: (2025)

Rethinking Multi-domain Generalization with A General Learning Objective
par: Tan, Zhaorui, et autres
Publié: (2024)

NeuroGenPoisoning: Neuron-Guided Attacks on Retrieval-Augmented Generation of LLM via Genetic Optimization of External Knowledge
par: Zhu, Hanyu, et autres
Publié: (2025)

CTIArena: Benchmarking LLM Knowledge and Reasoning Across Heterogeneous Cyber Threat Intelligence
par: Cheng, Yutong, et autres
Publié: (2025)

Bleeding Pathways: Vanishing Discriminability in LLM Hidden States Fuels Jailbreak Attacks
par: Zhang, Yingjie, et autres
Publié: (2025)

IRCAN: Mitigating Knowledge Conflicts in LLM Generation via Identifying and Reweighting Context-Aware Neurons
par: Shi, Dan, et autres
Publié: (2024)

MEEA: Mere Exposure Effect-Driven Confrontational Optimization for LLM Jailbreaking
par: Zhang, Jianyi, et autres
Publié: (2025)

Reasoned Safety Alignment: Ensuring Jailbreak Defense via Answer-Then-Check
par: Cao, Chentao, et autres
Publié: (2025)

Babel: Jailbreaking Safety Attention via Obfuscation Distribution Optimized Sampling
par: Wang, Ziwei, et autres
Publié: (2026)

SafeDream: Safety World Model for Proactive Early Jailbreak Detection
par: Yan, Bo, et autres
Publié: (2026)

The Art of the Jailbreak: Formulating Jailbreak Attacks for LLM Security Beyond Binary Scoring
par: Hossain, Ismail, et autres
Publié: (2026)

LLM Jailbreak Detection for (Almost) Free!
par: Chen, Guorui, et autres
Publié: (2025)

Cracking Factual Knowledge: A Comprehensive Analysis of Degenerate Knowledge Neurons in Large Language Models
par: Chen, Yuheng, et autres
Publié: (2024)

On Jailbreaking Quantized Language Models Through Fault Injection Attacks
par: Zahran, Noureldin, et autres
Publié: (2025)

Automating Deception: Scalable Multi-Turn LLM Jailbreaks
par: Kumarappan, Adarsh, et autres
Publié: (2025)