AI Paper Atlas

Research map / Optimization / Implicit bias and generalization

Grokking phase transitions in learning local rules with gradient descent

Bojan Žunkovič, Enej Ilievski

JMLR 2024 · Optimization › Implicit bias and generalization · approach: Gradient flow and mean-field dynamics · 16 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add L1 regularisation instead of L2 (larger grokking probability, shorter grokking time, smaller effective dimension) (ablation)D-ball perceptron model · D-ball model, grokking probabilitygrokking probability: above 90% and always larger than 0.915 for L1; exponentially small in D for lambda1=0Improved
increase L2 regularisation strength lambda2 (ablation)D-ball perceptron model · D-ball model, grokking probabilitygrokking probability: increases up to a maximum depending on other parametersImproved
increase latent dimension D (L2 only) (ablation)D-ball perceptron model · D-ball model, grokking probabilitygrokking probability: exponentially decreases with DHurt
increase number of training samples N and class separation epsilon (ablation)1D exponential model · 1D model, grokking probability and timegrokking probability / grokking time: N=2, 5, 10 and epsilon_lambda=0.4, 0.04, 0.004 comparedImproved
add L2 regularisation to fixed-attention model (ablation)Tensor-network attention (fixed A) · Fixed attention, rule 30 (M=3 samples)grokking probabilityHurt
add small L1 regularisation to fixed-attention model (ablation)Tensor-network attention (fixed A) · Fixed attention, rule 30 (M=3 samples)grokking probabilityMixed
add L1 or L2 regularisation (ablation)Tensor-network attention (fixed A) · Fixed attention, rule 30, Examples 1-3critical exponent nu: nu 0.85, 0.75, 0.84 all below oneNo clear effect
add regularisation (L1 or L2) to full trained model (ablation)Tensor-network attention (trained A,B) · Full model, rule 30critical exponent nu: 2.5 -> 0.7 (L1), 0.9 (L2)Improved
add regularisation to full trained model (ablation)Tensor-network attention (trained A,B) · Full model, rule 30test error after grokkingHurt
increase bond dimension d of the model (ablation)Tensor-network attention (trained A,B) · Full model, rule 30critical exponent nu / effective dimensionNo clear effect
increase L2 regularisation in fixed attention model (ablation)Tensor-network attention (fixed A) · Fixed attention, rule 30grokking timeHurt
larger effective dimension D_eff and larger class separation epsilon (ablation)Tensor-network attention (fixed A) · Fixed attention, rule 30grokking time: larger D_eff gives longer grokking times; larger epsilon gives smallerMixed

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy