arXiv 2026-08-26 EN EVOMAL: Self-Poisoning in Self-Evolving Coding Agents Xiaodong Wu · Yu Shi · Qi Li · Zhimin Zhao +4
arXiv 2026-08-26 EN MeMark: Membrane-Space Watermarking for Spiking Neural Networks Roberto Riaño · Gorka Abad · Stjepan Picek · Aitor Urbieta
arXiv 2026-08-26 EN Pointing the Way, Hiding the Destination: Practical Private Dense Retrieval at Scale Peichun Hua · Danyang Chen · Junan Zhang · Haifeng Sun +4
arXiv 2026-08-26 EN Are LLM-Enhanced GNNs Privacy-Safe? Longzhu He · Zelang Wen · Chaozhuo Li · Sen Su
arXiv 2026-08-26 EN Adversarial Training of Linear Models under Stealthy Attacks Lovisa Eriksson · Dave Zachariah · André M. H. Teixeira
arXiv 2026-08-26 EN AI Slop and Hallucinations in Vulnerability Assessment: A Survey on Reasoning Failures and Trustworthy Mitigation Junchen Ding · Jialiang Dong · Yichen Zhu · Yi Liu +4
arXiv 2026-08-26 EN MMJailBench: A Factorized Benchmark for Disentangling Multimodal Jailbreak Vulnerabilities Tianshi Wang · Jingsong Wang · Yafei Huang · Fengling Li +2
arXiv 2026-08-26 EN MACGen: Toward Functionally Correct and Secure Code Generation via Multi-Agent Collaboration Miseon Yu · Jaehoon Choi · Younghan Lee · Yunheung Paek
arXiv 2026-08-26 EN Refusal geometry reflects refusal training: diverse refusal prefixes can raise stable rank and weaken refusal vector ablation attacks Andrey Labunets
arXiv 2026-08-26 EN LLMscope: Extracting LLM Assets from Edge AI Chips via Optical Probing Dev Mehta · Lily Dukette · William Folan · Olivia Kochol +3
