5 papers
Latent-space Attacks for Refusal Evasion in Language Models
Giorgio Piras, Raffaele Mura, Fabio Brau +4
Safety-aligned language models are trained to refuse harmful requests, yet refusal behavior can be suppressed by steering their internal representations. Existing methods do so by…
HO-FMN: Hyperparameter Optimization for Fast Minimum-Norm Attacks
Raffaele Mura, Giuseppe Floris, Luca Scionis +6
Gradient-based attacks are a primary tool to evaluate robustness of machine-learning models. However, many attacks tend to provide overly-optimistic evaluations as they use fixed l…
SOM Directions are Better than One: Multi-Directional Refusal Suppression in Language Models
Giorgio Piras, Raffaele Mura, Fabio Brau +3
Refusal refers to the functional behavior enabling safety-aligned language models to reject harmful or unethical prompts. Following the growing scientific interest in mechanistic i…
LatentBreak: Jailbreaking Large Language Models through Latent Space Feedback
Raffaele Mura, Giorgio Piras, KamilÄ LukoÅ¡iÅ«tÄ +3
Jailbreaks are adversarial attacks designed to bypass the built-in safety mechanisms of large language models. Automated jailbreaks typically optimize an adversarial suffix or adap…
Extracting Memorized Training Data via Decomposition
Ellen Su, Anu Vellore, Amy Chang +4
The widespread use of Large Language Models (LLMs) in society creates new information security challenges for developers, organizations, and end-users alike. LLMs are trained on la…