List your company

Modular Addition on Modular Addition f_train=0.5 P=97

2,460Grokking Onset (epochs)Norm Penalty
1,936.85,468.49,00012,531.6Jun 30, 2026

Evaluation results

MethodLinks
Norm PenaltyArchitecture=MLP, Regularization=Activation Norm Penalty2026.062,4606.3
Norm PenaltyArchitecture=Transformer, Regularization=Activation Norm Penalty2026.065,2001.5
Strong WD BaselineArchitecture=Transformer, Regularization=Weight Decay (10^-1)2026.068,000—
Strong WD BaselineArchitecture=MLP, Regularization=Weight Decay (10^-1)2026.0615,540—