Manuel Muñoz Plá PRO

mechanistic interpretability, attention heads, gauge invariance, identifiability, model pruning, vision transformers, language models, spherical codes, representation geometry, self-supervised learning, graph neural networks, medical imaging, reproducibility