Avery Griffin

neural networks · loss landscape degeneracy · effective parameter count interpretability · mechanistic interpretability · reverse engineering neural networks · sparse representations · dimensionality reduction · low-rank approximations interaction graphs · integrated gradients · modularity analysis