LLM Security

LLM: Large Language Model

Follow this topic Ask about this topic

Momentum

25 papers in the last four weeks, up 127% on the four weeks before. 0.2% of all new papers.

Jul 13Week of Sep 28

Latest papers 200

All topics

SecDrift: Measuring Sector-Conditioned Security Drift in AI-Generated Code
Jul 28, 2026Narayanaswami Natraj Bharadwaj, Dhivya ChandramouleeswaranCybersecurity LLM Security
Independent Researcher USA
The Illusion of Secure LLM Code: Closing the Security Gap via Iterative Reprompting
Jul 26, 2026Ishpuneet Singh, Shreyas Mahajan, Gurjot Singh +1LLM Prompting Cybersecurity
Department of Computer Science and Engineering, Thapar Institute of Engineering and Technology, Patiala, India · David R. Cheriton School of Computer Science, University of Waterloo, Waterloo, Canada
HiTMS: A High-Throughput Multi-Stream Linguistic Steganography Framework
Jul 26, 2026Ruiyi Yan, Zhongliang Yang, Yugo MurawakiLLM Security
Graduate School of Informatics, Kyoto University, Kyoto 606-8501, Japan · School of Cyberspace Security, Beijing University of Posts and Telecommunications, Beijing 100876, China
Language Shapes Instruction Hierarchy Compliance in Multilingual LLMs
Jul 26, 2026Jiwon Moon, Yerin Hwang, Kyomin JungMultilingual Language Model Evaluation LLM Security
IPAI, Seoul National University · Dept. of ECE, Seoul National University
Poster: Rethinking Security in LLM Code Generation through Real-World Risk Scenarios
Jul 25, 2026Lixun Ma, Ruolong Ma, Bei Wang +4LLM Safety Benchmarks LLM Security
Zhejiang University, Hangzhou, China · University College London, London, United Kingdom · Artificial Intelligence Institute, CAICT, Beijing, China +1
Leaky Language Models: Stealing Architecture and Inference Optimizations via Per-Token Timing
Jul 22, 2026Sadegh Majidi, Niloofar Mireshghallah, Kazem TaramModel Extraction Attacks Language Model Fingerprinting
Purdue University West Lafayette, IN, USA · Carnegie Mellon University Pittsburgh, PA, USA
HijackKV: New Threat in Position-Independent KV Cache Reuse
Jul 22, 2026Yichi Zhang, Zhiqi Wang, Huan Zhang +1KV Caching LLM Security
The Pennsylvania State University · University of Illinois Urbana-Champaign
Defense Against LLM Backdoors using Critical Neuron Isolation Pruning
Jul 22, 2026Yuxi Li, Zhibo Zhang, Kailong Wang +3LLM Backdoor Attacks Backdoor Defense in LLMs
Huazhong University of Science and Technology China · Nanjing University of Aeronautics and Astronautics China · Nanyang Technological University Singapore
Integrity of peer-to-peer distributed LLM inference under malicious nodes
Jul 21, 2026Mert Cihangiroglu, Antonino NoceraLLM Security Distributed Inference
DCALab, Department of Electrical, Computer and Biomedical Engineering, University of Pavia, Italy
CPInj: Uncovering Prompt Injection Risks in Textual Collaborative Prompt Optimization
Jul 21, 2026Xinting Liao, Behnoosh Zamanlooy, Masoumeh Shafieinejad +4LLM Security Prompt Injection Defense
University of British Columbia · 2Vector Institute · 3McMaster University +1
An Early Warning of Emerging Biosecurity Risks in Frontier LLMs
Jul 20, 2026Zhida He, Xia Hu, Baichen Le +20Language Model Safety Evaluation LLM Security
Shanghai Artificial Intelligence Laboratory
How Do You Choose Your AI Component? An Interview Study of Secure AI Integration in Practice
Jul 18, 2026Mahzabin Tamanna, Elizabeth Lin, Sparsha Gowda +2LLM Security Software Supply Chain Security
North Carolina State University, USA
Securing LLMs in the Wild: Privacy and Security Challenges at the Edge
Jul 13, 2026Ren-Yi Huang, Mingchen Li, Dumindu Samaraweera +1LLM Inference Efficiency Edge Computing
Department of Electrical and Computer Engineering, University of South Florida · Department of Mathematics, Embry-Riddle Aeronautical University
Large Language Models in Misinformation Ecosystems: Misuse, Defense, and Vulnerability
Jul 11, 2026Lingwei Wei, Dou Hu, Wei Zhou +2Fake News Detection Adversarial Attacks on LLMs
Institute of Information Engineering, Chinese Academy of Sciences, China · University of Illinois Chicago, USA · State Key Laboratory of Media Convergence and Communication, Communication University of China, China +1
Multi-Agent Firewall Architecture for Privacy Protection of Sensitive Data in Interactions with Language Models
Jul 9, 2026Hugo García Cuesta, Pablo Mateo Torrejón, Alfonso Sánchez-MaciánPrivacy Leakage in Language Models LLM Security
Universidad Carlos III de Madrid Madrid, Spain
Large Language Models (LLMs) and Generative AI in Cybersecurity and Privacy: A Survey of Dual-Use Risks, AI-Generated Malware, Explainability, and Defensive Strategies
Jul 8, 2026Kiarash Ahi, Saeed ValizadehLLM Security LLMs for Cybersecurity
Virelya Intelligence Research Labs, San Francisco Bay Area, California · Google, Mountain View, California
Beyond Static Rules: Automated Discovery of Latent Vulnerabilities in Text-to-SQL
Jul 4, 2026Hanqing Wang, Yongdong Chi, Jian Yang +4LLM Security Text-to-SQL
Shanghai University of Finance and Economics · Beihang University · Deepexi Technology Co. Ltd. +1
Black-Box Inference of LLM Architectural Properties with Restrictive API Access
Jul 1, 2026Christopher Ellis, Shreyas Chaudhari, Mei-Yu Wang +3Model Extraction Attacks LLM Security
Carnegie Mellon University Pittsburgh, PA 15213 · Pittsburgh Supercomputing Center
Beyond the Prompt: Jailbreaking Function-Calling LLMs via Simulated Moderation Traces
Jul 1, 2026Junlong Liu, Haobo Wang, Weiqi Luo +1LLM Security LLM Jailbreak Attacks
1Sun Yat-sen University · 2Nanyang Technological University
A Lifecycle and Application-Stack Survey of Large Language Model Vulnerabilities: Attacks, Risks, Defenses, and Open Problems
Jun 30, 2026Seyed Bagher Hashemi Natanzi, Bo TangAI Agent Security LLM Security
Department of Electrical and Computer Engineering, Worcester Polytechnic Institute, Worcester, MA, USA
CSO-LLM: Class Subspace Orthogonalization for Post-Training Backdoor Detection and Trigger Inversion in LLMs
Jun 30, 2026Zhengxing Li, David J. Miller, Guangmingmei Yang +1LLM Backdoor Attacks Backdoor Detection
Electrical Engineering Dept Penn State University University Park, PA 16802 · Anomalee Inc. · Computer Science & Engineering Dept Penn State University University Park, PA 16802 +1
An Empirical Study of Security Calibration in Large Language Models for Code
Jun 30, 2026Mohammed Latif Siddiq, Md. Nafiu Rahman, Joanna C. S. SantosLLM Security Language Model Calibration
University of Notre Dame Notre Dame, IN, USA · Brac University Dhaka, Bangladesh
Curvature-Guided Module Localization for Low-Rank Detoxification of Backdoored Large Language Models
Jun 29, 2026Arash Raftari, Mehrdad Mahdavi, Nathan Blackthorn +1LLM Backdoor Attacks Backdoor Defense in LLMs
AIVault Inc. Orlando, USA
On the Vulnerability of Parameter-Level Defenses to Model Merging
Jun 29, 2026Kuangpu Guo, Qingyan Zheng, Jian Liang +4LLM Security Language Model Merging
University of Science and Technology of China · School of Artificial Intelligence, University of Chinese Academy of Sciences · NLPR & MAIS, Institute of Automation, Chinese Academy of Sciences +1
SurrogateShield: Beyond Redaction for High-Utility, Privacy-Preserving LLM Interactions
Jun 28, 2026Sherwin Vishesh JathannaPrivacy-Preserving Language Models LLM Security
Arizona State University Tempe, Arizona, USA
LLM Semantic Signaling Game and Mechanism Design: Systematic Blindness, Awareness Shaping, and Mindset Dynamics
Jun 27, 2026Quanyan ZhuDeception in Language Models LLM Security
Department of Electrical and Computer Engineering New York University Tandon School of Engineering Brooklyn, NY 11201, USA
FlipGuard: Defending Large Language Models Against Quantization-Conditioned Backdoor Attacks
Jun 27, 2026Aoying Zheng, Anqi Du, Zizhuang Deng +1LLM Quantization LLM Backdoor Attacks
School of Cyber Science and Technology Shandong University · Suzhou Research Institute of Shandong University
On the Inseparability of Instructions and Data in Shared-Embedding Sequence Models
Jun 25, 2026Dewank Pant, Shruti Lohani, Avijit KumarLLM Security Prompt Injection Defense
Independent Researcher
ShareLock: A Stealthy Multi-Tool Threshold Poisoning Attack Against MCP
Jun 25, 2026Liwei Liu, Tianzhu Han, Zijian Liu +2Adversarial Attacks MCP Security
Shanghai Jiao Tong University · Shanghai, China
What Does It Mean to Break a Distillation Defense?
Jun 23, 2026Lena Libon, Pura Peetathawatchai, Michael Aerni +2Model Extraction Attacks LLM Security
ETH Zurich

Jul 28, 2026, cs.CRJ/K move · Enter open · S save

SecDrift: Measuring Sector-Conditioned Security Drift in AI-Generated Code

Narayanaswami Natraj Bharadwaj, Dhivya Chandramouleeswaran

Independent Researcher USA

Cybersecurity LLM Security Code Generation Evaluation Secure Code Generation LLM Code Generation

Open paper AskPDF arXiv

LLMs are increasingly used for code generation in critical infrastructure, yet the security effect of domain-specific prompting is understudied. We present SecDrift, a benchmark measuring sector-conditioned security drift: the change in static-analysis vulnerability rates when prompts are conditioned on industry contexts versus neutral baselines. We evaluate 7 LLMs (6 producing analyzable code) across 8 CISA critical infrastructure sectors and 9 CWE categories with 5 replicates (5,355 evaluations), using a 5-dimension transformation with a matched-baseline condition that holds the task fixed while substituting only domain terminology. Industry prompts naively appear more secure (14.0% vs. 11.4%, -2.7pp), but the gap is not statistically significant (Fisher's exact p = 0.24, Cohen's h = -0.08) and is a composition artifact of two CWE categories: excluding CWE-502 and CWE-22 eliminates and slightly reverses it (+0.4pp, p = 1.00). A mixed-effects logistic regression confirms sector identity is not a moderator and localizes the only detectable condition effect to those two vulnerability types. 0 of 8 sectors show drift distinguishable from baseline, corrected or uncorrected (|h| < 0.15). A placebo on two non-CISA sectors (e-commerce, online education) reproduces the CISA industry rate almost exactly (10.5% vs. 11.4%, p = 0.63): the small pooled pattern reflects generic industry-framing specificity, not critical-infrastructure identity. In contrast, model selection has a large and consistent effect: among full-output models vulnerability rates range from 11.6% to 16.1%, and these differences persist across conditions. Model choice, not prompt framing, is the more reliable security lever. We release the framework, prompts, generated code, findings, human-validation verdicts, and analysis scripts.

Momentum

Related topics

Latest papers 200