arXiv:2603.28964v2 Announce Type: replace
Abstract: We develop the spectral edge thesis: phase transitions in neural network training — grokking, capability gains, loss plateaus — are controlled by the spectral gap of the rolling-window Gram matrix …
arXiv:2602.16746v3 Announce Type: replace
Abstract: Grokking — the delayed transition from memorization to generalization in small algorithmic tasks — remains poorly understood. We present a geometric analysis of optimization dynamics in transformer…
arXiv:2602.18523v3 Announce Type: replace
Abstract: Grokking — the abrupt transition from memorization to generalization long after near-zero training loss — has been studied mainly in single-task settings. We extend geometric analysis to multi-task…
arXiv:2602.16967v3 Announce Type: replace
Abstract: Grokking — the abrupt transition from memorization to generalization after prolonged training — has been linked to confinement on low-dimensional execution manifolds in modular arithmetic. Whether …
arXiv:2603.28964v1 Announce Type: new
Abstract: We develop the spectral edge thesis: phase transitions in neural network training — grokking, capability gains, loss plateaus — are controlled by the spectral gap of the rolling-window Gram matrix of p…