The paper reports that dropout-based variance, embedding distribution shape, and neuron sparsity all shift around the moment a modular arithmetic network groks, and proposes these as forecasting signals.
Controlling Grokking with Nonlinearity and Data Symmetry
1 Pith paper cite this work. Polarity classification is still indexing.
abstract
This paper demonstrates that grokking behavior in modular arithmetic with a modulus P in a neural network can be controlled by modifying the profile of the activation function as well as the depth and width of the model. Plotting the even PCA projections of the weights of the last NN layer against their odd projections further yields patterns which become significantly more uniform when the nonlinearity is increased by incrementing the number of layers. These patterns can be employed to factor P when P is nonprime. Finally, a metric for the generalization ability of the network is inferred from the entropy of the layer weights while the degree of nonlinearity is related to correlations between the local entropy of the weights of the neurons in the final layer.
citation-role summary
citation-polarity summary
fields
cs.LG 1years
2025 1verdicts
REJECT 1roles
background 1polarities
unclear 1representative citing papers
citing papers explorer
-
Tracing the Path to Grokking: Embeddings, Dropout, and Network Activation
The paper reports that dropout-based variance, embedding distribution shape, and neuron sparsity all shift around the moment a modular arithmetic network groks, and proposes these as forecasting signals.