3 papers
cs.CL2026
LACUNA: A Testbed for Evaluating Localization Precision for LLM Unlearning
Matteo Boglioni, Thibault Rousset, Siva Reddy +2
LLMs memorize sensitive training data, including personally identifiable information (PII), creating a pressing need for reliable post hoc removal methods. Unlearning has emerged a…
cs.CL2025
Do Generalisation Results Generalise?
Matteo Boglioni, Andrea Sgobbi, Gabriel Tavernini +3
A large language model's (LLM's) out-of-distribution (OOD) generalisation ability is crucial to its deployment. Previous work assessing LLMs' generalisation performance, however, t…
cs.LG2025
Enhancing One-run Privacy Auditing with Quantile Regression-Based Membership Inference
Terrance Liu, Matteo Boglioni, Yiwei Fu +3
Differential privacy (DP) auditing aims to provide empirical lower bounds on the privacy guarantees of DP mechanisms like DP-SGD. While some existing techniques require many traini…