Multi-turn prompt injection attacks distribute malicious intent across multiple conversation turns, exploiting the assumption that each turn is evaluated independently. While single-turn detection has been extensively studied, no published formula exists for aggregating per-turn pattern scores into a conversation-level risk score at the proxy layer -- without invoking an LLM. We identify a fundamental flaw in the intuitive weighted-average approach: it converges to the per-turn score regardless of turn count, meaning a 20-turn persistent attack scores identically to a single suspicious turn. Drawing on analogies from change-point detection (CUSUM), Bayesian belief updating, and security risk-based alerting, we propose peak + accumulation scoring -- a formula combining peak single-turn risk, persistence ratio, and category diversity. Evaluated on 10,654 multi-turn conversations -- 588 attacks sourced from WildJailbreak adversarial prompts and 10,066 benign conversations from WildChat -- the formula achieves 90.8% recall at 1.20% false positive rate with an F1 of 85.9%. A sensitivity analysis over the persistence parameter reveals a phase transition at rho ~ 0.4, where recall jumps 12 percentage points with negligible FPR increase. We release the scoring algorithm, pattern library, and evaluation harness as open source.
Sentinel: Decoding Context Utilization via Attention Probing for Efficient LLM Context Compression
Ping An Technology (Shenzhen) Co., Ltd., China(平安科技(深圳)有限公司,中国);University of Science and Technology of China(中国科学技术大学);University of Electronic Science and Technology of China(电子科技大学)
Stability-Plasticity Balance via Singular-Vector Selection in LLM Continual Learning
School of Artificial Intelligence, Beihang University(北京航空航天大学人工智能学院);Beijing Advanced Innovation Center for Future Blockchain and Privacy Computing(未来区块链与隐私计算北京高精尖创新中心);Institute of Computing Technology, Chinese Academy of Sciences(中国科学院计算技术研究所)
Can LLMs Accurately Score Medical Diagnoses and Clinical Reasoning?
Wits MIND Institute, University of the Witwatersrand, Johannesburg, South Africa(维特士心理研究所,沃斯兰德大学,约翰内斯堡,南非);Grai Labs, Cape Town, South Africa(格雷实验室,开普敦,南非);South African Medical Research Council Vaccines and Infectious Diseases Analytics Research Unit, Faculty of Health Sciences, University of the Witwatersrand, Johannesburg, South Africa(南非医学研究理事会疫苗和传染病分析研究组,健康科学学院,沃斯兰德大学,约翰内斯堡,南非);Department of Internal Medicine, Charlotte Maxeke Johannesburg Academic Hospital, and Faculty of Health Sciences, University of the Witwatersrand, Johannesburg, South Africa(内科学系,查理·马克斯凯约翰内斯堡学术医院,以及健康科学学院,沃斯兰德大学,约翰内斯堡,南非);Department of Paediatrics and Child Health, Faculty of Health Sciences, University of the Witwatersrand, Johannesburg, South Africa(儿科学与儿童健康系,健康科学学院,沃斯兰德大学,约翰内斯堡,南非);Wits MIND Institute, University of the Witwatersrand, Johannesbu(维特士心理研究所,沃斯兰德大学,约翰内斯堡)
EvoLattice: Persistent Internal-Population Evolution through Multi-Alternative Quality-Diversity Graph Representations for LLM-Guided Program Discovery