Large Language Models (LLMs) are increasingly deployed in high-stakes settings, where they face diverse risks. Numerous defense strategies have been proposed to mitigate these risks, but they are almost always evaluated in isolation. This isolated view leaves a critical question open: does mitigating one risk inadvertently change a model's exposure to others? Beyond the well-studied risk-utility trade-off, we present the first systematic study of cross-risk interactions induced by LLM defenses. We propose CrossRiskEval, an evaluation paradigm that situates a defended model in a multi-dimensional risk space and quantifies how a defense built for one risk shifts the others. Among 166 cross-risk evaluations covering 32 defended models, 77.1% exhibit statistically significant cross-risk interactions. Most of these interactions amplify non-target risks, with increases exceeding 100% in some cases. Beyond behavioral evaluation, we conduct neuron-level analyses in seven selected cases to investigate one possible pathway associated with these interactions. We identify conflict-entangled neurons whose activation interventions produce opposing effects on proxies for the target and non-target risks. In conflict cases, restoring these neurons to their base-model activations partially reduces the corresponding risk increases, providing evidence that defense-induced changes to these neurons may contribute to the observed interactions. Building on this evidence, we propose Conflict-Aware Freezing, a training-time strategy that prevents direct updates to the parameters associated with the identified neurons. Across five conflict cases, it offsets 35%-196% of non-target risk amplification while meeting the original defense criterion.
Breaking Down the Defenses: A Comparative Survey of Attacks on Large Language Models
University of Illinois Urbana-Champaign(伊利诺伊大学厄巴纳-香槟分校);University of California, Los Angeles(加利福尼亚大学洛杉矶分校);University of Texas at Arlington(德克萨斯大学阿灵顿分校);Georgia Institute of Technology(佐治亚理工学院);Stanford University(斯坦福大学);Amazon GenAI(亚马逊生成式人工智能部门)
Toward Trustworthy AI: Multi-Target Adversarial Attacks and Robust Defenses for Continuous Data Summarization
Nankai University(南开大学);James Cook University(詹姆斯库克大学);Western Sydney University(西悉尼大学);Beijing University of Technology(北京工业大学);Fuzhou University(福州大学);Nanjing University of Science and Technology(南京理工大学);CSIRO's Data 61(澳大利亚联邦科学与工业研究组织Data61);The University of Adelaide(阿德莱德大学)
When Agents Learn to Be You: Benchmarking Privacy Leakage, Impersonation Risk, and Defenses in Persona Skills
The University of Sydney(悉尼大学);University of Queensland(昆士兰大学);Southeast University(东南大学);Mohamed bin Zayed University of Artificial Intelligence(穆罕默德·本·扎耶德人工智能大学)
Security in the Fine-Tuning Lifecycle of Large Language Models: Threats, Defenses,Evaluation, and Future Directions
Hangzhou Normal University(杭州师范大学);Zhejiang University(浙江大学);China Mobile (Zhejiang) Innovation Research Institute Co., Ltd.(中国移动(浙江)创新研究院有限公司);Quantum Cloud Computing and Distributed Systems (qCLOUDS) Lab, School of Computing and Information Systems(量子云计算与分布式系统(qCLOUDS)实验室,计算与信息学院);The University of Melbourne(墨尔本大学)
ChannelGuard: From Borrowed Safety to Application-Controlled Defenses in Multi-Agent LLMs
College of Engineering and Computer Science, University of Central Florida(工程与计算机科学学院,中央佛罗里达大学);Department of Computer Science and Engineering, North South University(计算机科学与工程系,北南大学);Computer Science and Engineering, Ahsanullah University of Science and Technology(计算机科学与工程,阿沙努拉科学与技术大学);Department of Electrical and Computer Engineering, Purdue University Fort Wayne(电气与计算机工程系,普渡大学弗拉特沃恩分校)
MemPoison: Uncovering Persistent Memory Threats and Structural Blind Spots in LLM Agents
State Key Laboratory for Novel Software Technology, Nanjing University, China(南京大学新型软件技术国家重点实验室);NARI Group Corporation/State Grid Electric Power Research Institute, China(国网电力研究院/国电电力集团)