Grokking as a First Order Phase Transition in Two Layer Networks

arXiv:2310.03789v3 Announce Type: replace-cross
Abstract: A key property of deep neural networks (DNNs) is their ability to learn new features during training. This intriguing aspect of deep learning stands out most clearly in recently reported Grokking phenomena. While mainly reflected as a sudden increase in test accuracy, Grokking is also believed to be a beyond lazy-learning/Gaussian Process (GP) phenomenon involving feature learning. Here we apply a recent development in the theory of feature learning, the adaptive kernel approach, to two …

abstract accuracy arxiv cond-mat.dis-nn cs.lg deep learning features key layer learn networks neural networks new features property stat.ml test training transition type while