4 papers
Universal Jailbreak Suffixes Are Strong Attention Hijackers
Matan Ben-Tov, Mor Geva, Mahmood Sharif
We study suffix-based jailbreaks$\unicode{x2013}$a powerful family of attacks against large language models (LLMs) that optimize adversarial suffixes to circumvent safety alignment…
GASLITEing the Retrieval: Exploring Vulnerabilities in Dense Embedding-based Search
Matan Ben-Tov, Mahmood Sharif
Dense embedding-based text retrieval$\unicode{x2013}$retrieval of relevant passages from corpora via deep learning encodings$\unicode{x2013}$has emerged as a powerful method attain…
Impactful Bit-Flip Search on Full-precision Models
Nadav Benedek, Matan Levy, Mahmood Sharif
Neural networks have shown remarkable performance in various tasks, yet they remain susceptible to subtle changes in their input or model parameters. One particularly impactful vul…
CaFA: Cost-aware, Feasible Attacks With Database Constraints Against Neural Tabular Classifiers
Matan Ben-Tov, Daniel Deutch, Nave Frost +1
This work presents CaFA, a system for Cost-aware Feasible Attacks for assessing the robustness of neural tabular classifiers against adversarial examples realizable in the problem…