Hi, I'm Thomas! I'm a first year PhD student working on mechanistic interpretability in machine learning.

I work under the supervision of Jose M. Oramas at the University of Antwerp within IDLab-imec. My research focusses on scaling up current mechanistic interpretability techniques to increasingly characterise current language models.

Machine learning models have become an undeniably useful tool in recent years. Yet, they remain shrouded in mystery; both in research and everyday use. Understanding these models will not only help their trustworthiness but also make it easier to improve them.

Outside of work, I love to go on long bike rides, game with friends and solve chess puzzles.