1.5k citations · 2.4k across the 38 of their papers we have counts for
8 papers · 1 filter
Deep-Fake CAPTCHA: Mitigating Next-Generation Social Engineering Attacks
Guy Frankovits, Lior Yasur, Fred M. Grabovski +1
This paper presents DF-CAPTCHA, an active defense against real-time deepfake impersonation in voice and video calls. Instead of passively searching for artifacts, DF-CAPTCHA prompt…
Detokenization Leaks: Reconstructing Local LLM Outputs From Cache Traces
Roy Weiss, Benyamin Konstantinov, Eitam Sheetrit +2
We present a new attack that reconstructs the text generated by locally hosted LLMs by observing CPU cache activity during detokenization. Unlike prior attacks that rely on deploym…
Know Your Agent: Reconnaissance-Driven Pentesting of AI Agents
Or Zion Eliav, Eyal Lenga, Shir Bernstien +1
Traditional pentesting uses reconnaissance at each step to uncover unseen weaknesses, build stronger attacks, and advance the objective; we argue that AI agents require the same tr…
PRISM: Recovering Instruction Sets from Language Model Activations
Gilad Gressel, Rahul Pankajakshan, Julia Diament +3
As LLMs are deployed as agents, reliable monitoring requires knowing not only what they output, but which instructions are steering their behavior. This is difficult when models in…
A Few Steps Further: Why Defenses Against Malicious Finetuning Erode Under Continued Training
Itay Zloczower, Eyal Lenga, Gilad Gressel +1
Model providers increasingly release the weights of large language models. Although these models are safety-aligned before release, their safeguards can often be removed by fine-tu…
Who Owns This Agent? Tracing AI Agents Back to Their Owners
Ruben Chocron, Doron Jonathan Ben Chayim, Eyal Lenga +3
AI agents increasingly act autonomously in the world, yet harmful behavior cannot be reliably traced to the account that deployed the agent. This creates an accountability gap acro…
LeakBoost: Perceptual-Loss-Based Membership Inference Attack
Amit Kravchik Taub, Fred M. Grabovski, Guy Amit +1
Membership inference attacks (MIAs) aim to determine whether a sample was part of a model's training set, posing serious privacy risks for modern machine-learning systems. Existing…
GAVEL: Towards Rule-Based Safety Through Activation Monitoring
Shir Rozenfeld, Rahul Pankajakshan, Itay Zloczower +3
Large language models (LLMs) are increasingly paired with activation-based monitoring to detect and prevent harmful behaviors that may not be apparent at the surface-text level. Ho…