two-layer-networks

This empirical study validates theoretical findings on feature repulsion and spectral lock-in during the grokking phenomenon in two-layer neural networks, demonstrating how activation functions influence the transition from memorization to generalization.