Cover Image

Uncovering Logit Suppression Vulnerabilities in LLM Safety Alignment

Fuente: arXiv

Saved in:

Bibliographic Details
Main Authors:	Li, Yuxi, Liu, Yi, Li, Yuekang, Shi, Ling, Deng, Gelei, Chen, Shengquan, Wang, Kailong
Format:	Preprint
Published:	2024
Subjects:	Cryptography and Security Artificial Intelligence Machine Learning
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

Continuous Embedding Attacks via Clipped Inputs in Jailbreaking Large Language Models
by: Xu, Zihao, et al.
Published: (2024)

A Comprehensive Study of Jailbreak Attack versus Defense for Large Language Models
by: Xu, Zihao, et al.
Published: (2024)

Model-Editing-Based Jailbreak against Safety-aligned Large Language Models
by: Li, Yuxi, et al.
Published: (2024)

A Rusty Link in the AI Supply Chain: Detecting Evil Configurations in Model Repositories
by: Ding, Ziqi, et al.
Published: (2025)

Turning Bias into Bugs: Bandit-Guided Style Manipulation Attacks on LLM Judges
by: Yang, Xianglin, et al.
Published: (2026)

Embedding Poisoning: Bypassing Safety Alignment via Embedding Semantic Shift
by: Yuan, Shuai, et al.
Published: (2025)

Uncovering, Explaining, and Mitigating the Superficial Safety of Backdoor Defense
by: Min, Rui, et al.
Published: (2024)

Safety Tax: Safety Alignment Makes Your Large Reasoning Models Less Reasonable
by: Huang, Tiansheng, et al.
Published: (2025)

Attention Eclipse: Manipulating Attention to Bypass LLM Safety-Alignment
by: Zaree, Pedram, et al.
Published: (2025)

Credential Leakage in LLM Agent Skills: A Large-Scale Empirical Study
by: Chen, Zhihao, et al.
Published: (2026)

SoSBench: Benchmarking Safety Alignment on Six Scientific Domains
by: Jiang, Fengqing, et al.
Published: (2025)

AlphaAlign: Incentivizing Safety Alignment with Extremely Simplified Reinforcement Learning
by: Zhang, Yi, et al.
Published: (2025)

Digger: Detecting Copyright Content Mis-usage in Large Language Model Training
by: Li, Haodong, et al.
Published: (2024)

Agent Skills in the Wild: An Empirical Study of Security Vulnerabilities at Scale
by: Liu, Yi, et al.
Published: (2026)

Pharmacist: Safety Alignment Data Curation for Large Language Models against Harmful Fine-tuning
by: Liu, Guozhi, et al.
Published: (2025)

Pandora: Jailbreak GPTs by Retrieval Augmented Generation Poisoning
by: Deng, Gelei, et al.
Published: (2024)

Supply-Chain Poisoning Attacks Against LLM Coding Agent Skill Ecosystems
by: Qu, Yubin, et al.
Published: (2026)

SAFEx: Analyzing Vulnerabilities of MoE-Based LLMs via Stable Safety-critical Expert Identification
by: Lai, Zhenglin, et al.
Published: (2025)

Safety Alignment Can Be Not Superficial With Explicit Safety Signals
by: Li, Jianwei, et al.
Published: (2025)

Syntax- and Compilation-Preserving Evasion of LLM Vulnerability Detectors
by: Sun, Luze, et al.
Published: (2026)

When Safety Geometry Collapses: Fine-Tuning Vulnerabilities in Agentic Guard Models
by: Hossain, Ismail, et al.
Published: (2026)

Prompt Injection attack against LLM-integrated Applications
by: Liu, Yi, et al.
Published: (2023)

Unsafe LLM-Based Search: Quantitative Analysis and Mitigation of Safety Risks in AI Web Search
by: Luo, Zeren, et al.
Published: (2025)

In Vino Veritas and Vulnerabilities: Examining LLM Safety via Drunk Language Inducement
by: Shetty, Anudeex, et al.
Published: (2026)

Efficient but Vulnerable: Benchmarking and Defending LLM Batch Prompting Attack
by: Yue, Murong, et al.
Published: (2025)

TombRaider: Entering the Vault of History to Jailbreak Large Language Models
by: Ding, Junchen, et al.
Published: (2025)

RASA: Routing-Aware Safety Alignment for Mixture-of-Experts Models
by: Liang, Jiacheng, et al.
Published: (2026)

SNARE: Adaptive Scenario Synthesis for Eliciting Overeager Behavior in Coding Agents
by: Qu, Yubin, et al.
Published: (2026)

Jailbreaking and Mitigation of Vulnerabilities in Large Language Models
by: Peng, Benji, et al.
Published: (2024)

Differentiated Directional Intervention A Framework for Evading LLM Safety Alignment
by: Zhang, Peng, et al.
Published: (2025)

Enhancing Model Defense Against Jailbreaks with Proactive Safety Reasoning
by: Yang, Xianglin, et al.
Published: (2025)

Logit-Gap Steering: A Forward-Pass Diagnostic for Alignment Robustness
by: Li, Tung-Ling, et al.
Published: (2025)

Immune: Improving Safety Against Jailbreaks in Multi-modal LLMs via Inference-Time Alignment
by: Ghosal, Soumya Suvra, et al.
Published: (2024)

Toward Principled LLM Safety Testing: Solving the Jailbreak Oracle Problem
by: Lin, Shuyi, et al.
Published: (2025)

Knowing without Acting: The Disentangled Geometry of Safety Mechanisms in Large Language Models
by: Wu, Jinman, et al.
Published: (2026)

Amplification Effects in Test-Time Reinforcement Learning: Safety and Reasoning Vulnerabilities
by: Khattar, Vanshaj, et al.
Published: (2026)

Improved Algorithms for Differentially Private Language Model Alignment
by: Chen, Keyu, et al.
Published: (2025)

Rethinking the Vulnerability of Concept Erasure and a New Method
by: Richardson, Alex D., et al.
Published: (2025)

Uncovering Vulnerabilities of LLM-Assisted Cyber Threat Intelligence
by: Meng, Yuqiao, et al.
Published: (2025)

"Do Not Mention This to the User": Detecting and Understanding Malicious Agent Skills
by: Liu, Yi, et al.
Published: (2026)