A novel methodology is proposed that explicitly balances the trade-off among detection performance, computational requirements, and robustness, and introduces system-level threat models that capture how attackers exploit different degrees of knowledge to evade the entire Compound AI System rather than isolated detector...
Andrea Ponte, Luca Demetrio, Luca Oneto et al.· 0 citations
A unified evaluation framework based on a comprehensive pool of minimum-norm attacks and robustness--perturbation curves across $\ell_0, $\ell_1, $\ell_2$ and $\ell_\infty$ norms", and proposes the Defense Optimality Index to rank defenses by their gap to the defense frontier.
ExE-Bench assesses performance, temporal and adversarial robustness, and computational overhead, aggregating them into a single score for direct and fair model comparison, and highlights how evaluations conducted only after deployment are suboptimal and unable to provide a complete picture of their performance.
Andrea Ponte, Daniel Gibert, M. Kozák et al.· arXiv.org· 0 citations
This paper introduces dormant steering logic into the model architecture through a trigger-gated additive modification of an intermediate representation, without poisoning training data, controlling downstream fine-tuning, or modifying prompts at deployment time.
M. Briglia, Igor Maljkovic, A. E. Cinà et al.· 0 citations
We use cookies to run the site and, with your consent, for analytics and to show ads.
See our Cookie Policy.
