2 papers
cs.AI2026
xMIx: High-Performance Serving-Time Platform for Mechanistic Interpretability Apps
Michael Blum, Mark Silberstein, Yaniv David
Mechanistic interpretability (MI) has emerged as a powerful approach for analyzing and intervening in inference computations, with a growing number of applications such as jailbrea…
cs.CR2025
AMuLeT: Automated Design-Time Testing of Secure Speculation Countermeasures
Bo Fu, Leo Tenenbaum, David Adler +7
In recent years, several hardware-based countermeasures proposed to mitigate Spectre attacks have been shown to be insecure. To enable the development of effective secure speculati…