Conferences

At ICML 2024 in Vienna, we presented our paper on using mechanistic interpretability to extract nuclear-physics structure from neural networks trained on nuclear data.