My first paper at LG AI Research was accepted to ICML workshop on Mechanistic Interpretability! This is also my first paper in mechanistic interpretability, where I explore graph transformers and their ability to understand graph input and the composition they make to solve graph tasks. You can find the paper here.