Enterprise AI agents that share a memory store face two unaddressed risks: sensitive data can leak through legitimately computed results the requester could not derive, and departments can silently compute a same-named key performance indicator (KPI) through conflicting logic. Existing agent-memory systems (e.g., MemGP...
While open-weight models have enjoyed steady progress in capabilities and wide adoption across multiple domains, their safety remains an important concern. One key feature is the ability to refuse or deflect harmful, malicious, or insensitive prompts. In this paper, we expose safety vulnerabilities across six common op...
Abhinav Sudhakar Dubey (University of California Santa Cruz), Scott Sirri (University of California Santa Cruz), Vaggos Chatziafratis (University of California Santa Cruz) et al.· 0 citations
LLM agents increasingly act in multi-actor environments, exposing them to misuse from multiple sources: decomposition attacks, where a harmful request is split into innocuous sub-requests, and prompt injection attacks, where a compromised tool delivers a malicious instruction. Existing evaluations treat these threats s...
Aniruddh Pramod, James Oldfield, Adel Bibi· 0 citations
Image-to-text jailbreaks place harmful intent in text, image content, or the relationship between them. We examine image-side factors across four published attack families on a 313-prompt StrongREJECT slice, using five multimodal models and an additional appendix evaluation of InternVL3.5-8B. The harmful instruction is...
Boyuan Chen, Yehia Dawoud, Hailemariam Mersha et al.· 0 citations
Reach audiences
Advertise in front of researchers, engineers, and readers.
We audit frequency and decoder-depth claims for AdvWave-P, an additive audio jailbreak, on Qwen2-Audio. The protocol masks frequency components of the perturbation in the short-time Fourier transform (STFT) domain and measures attack success and audio-span representations. On 520 AdvBench prompts, the primary judge lab...
Boyuan Chen, Minseok Kim, Sohaila Abdulsattar et al.· 0 citations
Existing latent-based watermarking methods for diffusion models have overestimated their robustness to image distortions, including geometric transformations such as rotation, scaling, and translation (RST). Moreover, this paradigm of watermarking approaches may suffer from inherent limitations arising from the domain...
Cheng-Han Yeh, Kuan-chun Yu, Cheng-Chang Tsai et al.· 0 citations
In this paper, we show that standard evaluations of high-resolution Model Inversion Attacks (MIAs) significantly underestimate training-data privacy leakage. State-of-the-art privacy defenses, standard training techniques such as MixUp and Adversarial Training, and undefended models all leak training images at rates 1....
Shailen Smith, Rasmus Torp, Adam Breuer· 0 citations
Safety auditing of post-trained large language models typically relies on model behavior, requiring model execution and depending on the coverage of available evaluations. This work asks a different question: Do the target behaviors optimized during supervised fine-tuning (SFT) leave readable evidence directly in check...
Ziqun Bao, Xinyu Zhang, Yuchen Shao et al.· 0 citations
Interpretability methods for neural networks are predominantly reactive: they analyse activations produced during specific forward passes, requiring known inputs to find hidden capabilities such as backdoors. We propose Weight Oracles, fine-tuned language models that diagnose properties of a target network by reading i...
Krishna Kabra, Constantin Venhoff, Christian Schroeder de Witt· 0 citations
Security evaluation of learning-based systems requires more than just testing the system against a fixed collection of attacks. It requires adaptive mechanisms that can efficiently discover \textit{sets} of inputs that induce model failure. We introduce ATLAS (Adaptive Trust-Regions for Latent Adversarial Searches), wh...
Reinforcement learning with verifiable rewards (RLVR) trains a language model on problems that may themselves be confidential, and the trained model can reveal which problems it saw. We study RLVR under prompt-level differential privacy: the released weights must be ({\epsilon},{\delta})-differentially private with res...
Watermarking has been proposed as a lightweight mechanism to identify AI-generated text, with schemes typically relying on perturbations to token distributions. While prior work shows that paraphrasing can weaken such signals, these attacks remain partially detectable or degrade text quality. We demonstrate that cross-...
