The release decision for frontier AI systems increasingly relies on cyber capability benchmarks, yet public vulnerability benchmarks can expose agents to previously published advisories, exploits, and fixes, making it difficult to distinguish prior exposure from capability on unseen vulnerabilities. We evaluate open-we...
Tobias Heldt, Matthew Turk, Christoph R. Landolt et al.· 0 citations
IoT encompasses diverse physical entities, from smart home devices to autonomous vehicles, creating a complex environment with heterogeneous security models. This heterogeneity makes IoT sub-systems vulnerable to various network attacks. Modern security systems must therefore be more robust to ensure security and priva...
Tool-using language-model agents are vulnerable to indirect prompt injection because they must act on untrusted external content. Existing training-time defenses can reduce attack success rates, but often at the cost of general capabilities. We show that training-based defenses induce substantial drift in the model's o...
Mohamed Dhouib, Clément Elliker, Alexi Canesse et al.· 0 citations
Localizing where an LLM-based agent first fails to uphold security during a repair can show which stage of its workflow needs an additional safeguard. This is difficult for silent failures, which are patches that pass syntactic and functional checks but still contain a security vulnerability. Because such patches give...
Wenji Bai, Muhammad Waseem, Zeeshan Rasheed et al.· 0 citations
Reach audiences
Advertise in front of researchers, engineers, and readers.
The robustness of Personally Identifiable Information (PII) protection in Large Language Models (LLMs) is a critical concern, yet the risks associated with cross-lingual data extraction remain under-explored. This study evaluates the vulnerability of English-centric and multilingual models to Training Data Extraction (...
Alexandru Nazare, Agnese Profico, Nicol\`o Vania et al.· 0 citations
Sparse autoencoders (SAEs) are increasingly used not only to interpret language models but also to intervene on their internal representations. We show that this creates a supply-chain attack surface: a maliciously modified SAE can induce attacker-chosen behavior when inserted into the forward pass of an otherwise unch...
E. Ahlers, Daniel Passon, Tobias Kiecker et al.· 0 citations
Agent skills package task-specific knowledge and procedures that can be composed to support complex agent tasks, while public marketplaces provide a growing pool of reusable skills. Existing security vetting, however, largely evaluates skills in isolation, leaving composition-induced risks underexplored. Such risks ari...
Zunlong Zhou, Ziyuan Yang, Mengyu Sun et al.· 0 citations
We introduce a framework for multi-bit text watermarking with security defined directly through $f$-divergence from the base language model distribution. Unlike prior approaches that focus on average-key distortion-freeness or a particular statistical distance, our formulation supports general $f$-divergences, includin...
Benjamin D. Kim, Wan-Rong Zhang, Wei-Tong Ruan et al.· 0 citations
As Large Language Models (LLMs) gain tool use capabilities such as web search, they can retrieve and cross-reference public information, creating privacy risks beyond memorization. One manifestation is re-identification: linking an anonymized interview transcript to a named individual. Yet without ground-truth identiti...
Mechanistic interpretability has emerged as the primary means to understand safety behavior of LLMs. However, existing tools primarily focus on the activating neurons or features of a model. The role of the remaining large set of inactive components is invisible to such methods. This work demonstrates that the inactive...
The increasing complexity and scale of modern cyber-attacks demand intelligent and computationally efficient Intrusion Detection Systems (IDS). However, designing effective data pre-processing pipelines traditionally involves substantial trial-and-error effort and repeated evaluation of alternative configurations. For...
Bao-Phong Nguyen, G. Pham, Thai-Duong Do et al.· Journal of Supercomputing· 0 citations
Self-play red-teaming improves language-model safety by pitting attacker and defender roles against each other in a zero-sum game. However, real adversaries increasingly use mosaic attacks: multi-turn sequences whose individual fragments are innocuous in isolation yet assemble into a harmful payload. We develop a theor...
Emanuele La Malfa, Saar Cohen, Gabriele La Malfa et al.· 0 citations
