Safety evaluation for advanced AI systems assumes that behavior observed under evaluation predicts behavior in deployment. This assumption weakens for agents with situational awareness, which may exploit regime leakage, cues distinguishing evaluation from deployment, to implement conditional policies that comply under oversight while defecting in deployment-like regimes. We recast alignment evaluation as a problem of information flow under partial observability and show that divergence between evaluation-time and deployment-time behavior is bounded by the regime information extractable from decision-relevant internal representations. We study regime-blind mechanisms, training-time interventions that restrict access to regime cues through adversarial invariance constraints without assuming complete information erasure. We evaluate this approach across multiple open-weight language models and controlled failure modes including scientific sycophancy, temporal sleeper agents, and data leakage. Regime-blind training reduces regime-conditioned failures without measurable loss of task utility, but exhibits heterogeneous and model-dependent dynamics. Sycophancy shows a sharp representational and behavioral transition at moderate intervention strength, consistent with a stability cliff. In sleeper-style constructions and certain cross-model replications, suppression occurs without a clean collapse of regime decodability and may display non-monotone or oscillatory behavior as invariance pressure increases. These findings indicate that representational invariance is a meaningful but limited control lever. It can raise the cost of regime-conditioned strategies but cannot guarantee elimination or provide architecture-invariant thresholds. Behavioral evaluation should therefore be complemented with white-box diagnostics of regime awareness and internal information flow.
PRISM: Generation-Time Detection and Mitigation of Secret Leakage in Multi-Agent LLM Pipelines
School of Computing and Electrical Engineering(计算与电子工程学院);Indian Institute of Technology, Mandi (IIT)(印度理工学院,曼迪(IIT));The Alan Turing Institute, London(阿兰·图灵研究所,伦敦);University of Warwick(沃里克大学)
Agent Tools Orchestration Leaks More: Dataset, Benchmark, and Mitigation
Institute of Information Engineering, Chinese Academy of Sciences(中国科学院信息工程研究所);School of Cyber Security, University of Chinese Academy of Sciences(中国科学院大学网络空间安全学院)
AI Slop and Hallucinations in Vulnerability Assessment: A Survey on Reasoning Failures and Trustworthy Mitigation
University of New South Wales(新南威尔士大学);University of Wollongong(伍伦贡大学);Nanjing University of Information Science & Technology(南京信息工程大学);Griffith University(格里菲斯大学);Nanyang Technological University(南洋理工大学)
Risk Taxonomy, Mitigation, and Assessment Benchmarks of Large Language Model Systems
Zhongguancun Laboratory(中关村实验室);Tsinghua University(清华大学);Institute of Information Engineering, Chinese Academy of Sciences(中国科学院信息工程研究所);Ant Group(蚂蚁集团)
When Can We Trust Early Warnings? Leakage-Excluded Early Outcome Prediction from LMS Interaction Logs
Gamaizer;Université de technologie de Compiègne, CNRS, Heudiasyc(巴黎综合理工学院,法国国家科学研究中心,Heudiasyc实验室);Sorbonne Université, CNRS UMR 7585, LPMHE(索邦大学,法国国家科学研究中心,LPMHE实验室)