Grokking and Generalization Collapse: Insights from \texttt{HTSR} theory

• URL: http://arxiv.org/abs/2506.04434v1
• Date: Wed, 04 Jun 2025 20:34:37 GMT
• Title: Grokking and Generalization Collapse: Insights from \texttt{HTSR} theory
• Abstract summary: We study the well-known grokking phenomena in neural networks (NNs) using a 3-layer trained on 1 k-sample subset of MNIST.<n>We discover a novel third phase -- emphanti-grokking -- that occurs very late in training and resembles but is distinct from the familiar emphpre-grokking phases.
• Score: 4.371245698184159
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: We study the well-known grokking phenomena in neural networks (NNs) using a 3-layer MLP trained on 1 k-sample subset of MNIST, with and without weight decay, and discover a novel third phase -- \emph{anti-grokking} -- that occurs very late in training and resembles but is distinct from the familiar \emph{pre-grokking} phases: test accuracy collapses while training accuracy stays perfect. This late-stage collapse is distinct, from the known pre-grokking and grokking phases, and is not detected by other proposed grokking progress measures. Leveraging Heavy-Tailed Self-Regularization HTSR through the open-source WeightWatcher tool, we show that the HTSR layer quality metric $\alpha$ alone delineates all three phases, whereas the best competing metrics detect only the first two. The \emph{anti-grokking} is revealed by training for $10^7$ and is invariably heralded by $\alpha < 2$ and the appearance of \emph{Correlation Traps} -- outlier singular values in the randomized layer weight matrices that make the layer weight matrix atypical and signal overfitting of the training set. Such traps are verified by visual inspection of the layer-wise empirical spectral densities, and by using Kolmogorov--Smirnov tests on randomized spectra. Comparative metrics, including activation sparsity, absolute weight entropy, circuit complexity, and $l^2$ weight norms track pre-grokking and grokking but fail to distinguish grokking from anti-grokking. This discovery provides a way to measure overfitting and generalization collapse without direct access to the test data. These results strengthen the claim that the \emph{HTSR} $\alpha$ provides universal layer-convergence target at $\alpha \approx 2$ and underscore the value of using the HTSR alpha $(\alpha)$ metric as a measure of generalization.

Related papers

• Harmfulness Propagation Dynamics: Layer-wise Trajectories of Adversarial Intent in Large Language Models [0.0]
  textbfherald (textbfEncoding textbfRecognition via textbfActivation textbfLayer textbfDynamics)<n>textbfherald (textbfHarmful textbfEncoding textbfRecognition via textbfActivation textbfLayer textbfDynamics)<n>
  arXiv  Detail & Related papers  (2026-09-11T21:04:55Z)
• Demixing Sparse Signals from Nonlinear Observations using Generalized Non-convex Regularization [0.0]
  We consider the recovery of a pair of sparse vectors from a limited number of nonlinear observations of their superposition.<n>We propose a regularization-based framework combining a Huberized data fidelity with generalized folded-concave penalties.<n>We derive estimation error bounds of order $sqrtslog(n)/m$ that hold at emphevery localized stationary point.
  arXiv  Detail & Related papers  (2026-07-12T07:29:39Z)
• How Neural Reward Models Learn Features for Policy Optimization: A Single-Index Analysis [53.063298916923976]
  We study the feedback in a Gaussian single-index model with $r*(x) = *(langle *, xrangle)$ and $x sim N(0, I_d)$.<n>We analyze a two-stage neural reward model that first learns the hidden direction $*$ from reward-weighted samples and then fits the readout layer by weighted ridge regression.
  arXiv  Detail & Related papers  (2026-05-23T22:00:38Z)
• When Stronger Triggers Backfire: A High-Dimensional Theory of Backdoor Attacks [2.7998963147546143]
  Backdoor poisoning attacks behave counter-intuitively in high dimensions.<n>We study regularised generalised linear models on Gaussian-mixture data in the proportional regime.<n>Three phenomena emerge: (i) clean test accuracy increases with $$; (ii) attack success peaks at a finite $$ and then declines.
  arXiv  Detail & Related papers  (2026-05-21T13:39:06Z)
• Detecting overfitting in Neural Networks during long-horizon grokking using Random Matrix Theory [1.6615337656760856]
  We present a novel Random Matrix Theory method that detects the onset of overfitting in deep learning models without access to train or test data.<n>During the onset of overfitting, which we call the "anti-grokking" phase in long-horizon grokking, Correlation Traps form and grow in number and scale.<n>Our findings show that anti-grokking is an additional grokking phase with high train accuracy and decreasing test accuracy, structurally distinct from pre-grokking through its Correlation Traps.
  arXiv  Detail & Related papers  (2026-05-12T16:57:48Z)
• Tight Convergence Rates for Online Distributed Linear Estimation with Adversarial Measurements [66.94250413799232]
  We study mean estimation of a random vector $X$ in a distributed parameter-server-worker setup.<n>Key challenges are adversarial measurements and asynchrony.<n>Results provide a unified finite-time characterization of robustness, identifiability, and statistical efficiency in distributed linear estimation with adversarial workers.
  arXiv  Detail & Related papers  (2026-04-07T11:45:55Z)
• Late-Stage Generalization Collapse in Grokking: Detecting anti-grokking with Weightwatcher [1.6615337656760856]
  emphMemorization in neural networks lacks a precise operational definition and is often inferred from the grokking regime.<n>We identify a previously unreported third phase of grokking in this training regime: emphanti-grokking, a late-stage collapse of generalization.
  arXiv  Detail & Related papers  (2026-02-02T22:09:14Z)
• Singular Bayesian Neural Networks [1.2891210250935148]
  Bayesian neural networks promise calibrated uncertainty but require $O(mn)$ parameters for standard mean-field Gaussian posteriors.<n>We induce a posterior that is singular with respect to the Lebesgue measure, concentrating on the rank-$r$ manifold.<n>We derive PAC-Bayes generalization bounds whose complexity term scales as $sqrtr(m+n)$ instead of $sqrtm n$, and prove loss bounds that decompose the error into optimization and rank-induced bias.
  arXiv  Detail & Related papers  (2026-01-30T23:06:34Z)
• Revisiting Weighted Strategy for Non-stationary Parametric Bandits and MDPs [56.246783503873225]
  This paper revisits the weighted strategy for non-stationary parametric bandits.<n>We propose a simpler weight-based algorithm that is as efficient as window/restart-based algorithms.<n>Our framework can be used to improve regret bounds of other parametric bandits.
  arXiv  Detail & Related papers  (2026-01-03T04:50:21Z)
• Binary perceptron computational gap -- a parametric fl RDT view [2.538209532048867]
  asymmetric binary perceptron (ABP) likely exhibits two phase transitioning constraint density thresholds.<n>We consider a particular parametric utilization of emphfully lifted random duality theory (fl RDT) [85] and study its potential ABP's algorithmic implications.
  arXiv  Detail & Related papers  (2025-11-02T18:23:49Z)
• deep-REMAP: Probabilistic Parameterization of Stellar Spectra Using Regularized Multi-Task Learning [0.0]
  deep-REMAP is a novel deep learning framework that utilizes a regularized, multi-task approach to predict stellar atmospheric parameters from observed spectra.<n>We train a deep convolutional neural network on the PHOENIX synthetic spectral library and use transfer learning to fine-tune the model on a small subset of observed FGK dwarf spectra.<n>Deep-REMAP accurately recovers the effective temperature ($T_rmeff$), surface gravity ($log rmg$), and metallicity ([Fe/H]), achieving a precision of, for instance, approximately 75 K in $T
  arXiv  Detail & Related papers  (2025-10-10T13:20:06Z)
• Sharper Guarantees for Learning Neural Network Classifiers with Gradient Methods [43.32546195968771]
  We study the data-dependent convergence and generalization behavior of gradient methods for neural networks with smooth activation.<n>Our results improve upon the shortcomings of the well-established Rademacher complexity-based bounds.<n>We show that a large step-size significantly improves upon the NTK regime's results in classifying the XOR distribution.
  arXiv  Detail & Related papers  (2024-10-13T21:49:29Z)
• Constructive Universal Approximation and Finite Sample Memorization by Narrow Deep ReLU Networks [0.0]
  We show that any dataset with $N$ distinct points in $mathbbRd$ and $M$ output classes can be exactly classified.<n>We also prove a universal approximation theorem in $Lp(Omega; mathbbRm)$ for any bounded domain.<n>Our results offer a unified and interpretable framework connecting controllability, expressivity, and training dynamics in deep neural networks.
  arXiv  Detail & Related papers  (2024-09-10T14:31:21Z)
• Langevin dynamics for high-dimensional optimization: the case of multi-spiked tensor PCA [8.435118770300999]
  We show that the sample complexity required for recovering the spike associated with the largest SNR matches the well-known algorithmic threshold for the single-spike case.<n>As a key step, we provide a detailed characterization of the spikes and interactions that capture the high-dimensional trajectory dynamics.
  arXiv  Detail & Related papers  (2024-08-12T12:09:25Z)
• Learning with Norm Constrained, Over-parameterized, Two-layer Neural Networks [54.177130905659155]
  Recent studies show that a reproducing kernel Hilbert space (RKHS) is not a suitable space to model functions by neural networks. In this paper, we study a suitable function space for over- parameterized two-layer neural networks with bounded norms.
  arXiv  Detail & Related papers  (2024-04-29T15:04:07Z)
• Stochastic Gradient Succeeds for Bandits [64.17904367852563]
  We show that the emphstochastic gradient bandit algorithm converges to a emphglobally optimal policy at an $O (1/t)$ rate. Remarkably, global convergence of the gradient bandit algorithm has not been previously established.
  arXiv  Detail & Related papers  (2024-02-27T06:05:01Z)
• Local primordial non-Gaussianity from the large-scale clustering of photometric DESI luminous red galaxies [5.534428269834764]
  We use angular clustering of luminous red galaxies from the Dark Energy Spectroscopic Instrument (DESI) imaging surveys to constrain the local primordial non-Gaussianity parameter $fnl$. Our sample comprises over 12 million targets, covering 14,000 square degrees of the sky, with redshifts in the range 0.2 z 1.35$. We identify Galactic extinction, survey depth, and astronomical seeing as the primary sources of systematic error, and employ linear regression and artificial neural networks to alleviate non-cosmological excess clustering on large scales.
  arXiv  Detail & Related papers  (2023-07-04T14:49:23Z)
• MixCycle: Mixup Assisted Semi-Supervised 3D Single Object Tracking with Cycle Consistency [81.49482678098464]
  3D single object tracking (SOT) is an indispensable part of automated driving. Inspired by the great success of cycle tracking in unsupervised 2D SOT, we introduce the first semi-supervised approach to 3D SOT. Specifically, we introduce two cycle-consistency strategies for supervision: 1) Self tracking cycles, which leverage labels to help the model converge better in the early stages of training; 2) forward-backward cycles, which strengthen the tracker's robustness to motion variations and the template noise caused by the template update strategy.
  arXiv  Detail & Related papers  (2023-03-16T10:48:59Z)
• Revisiting Weighted Strategy for Non-stationary Parametric Bandits [82.1942459195896]
  This paper revisits the weighted strategy for non-stationary parametric bandits. We propose a refined analysis framework, which produces a simpler weight-based algorithm. Our new framework can be used to improve regret bounds of other parametric bandits.
  arXiv  Detail & Related papers  (2023-03-05T15:11:14Z)
• Improved techniques for deterministic l2 robustness [63.34032156196848]
  Training convolutional neural networks (CNNs) with a strict 1-Lipschitz constraint under the $l_2$ norm is useful for adversarial robustness, interpretable gradients and stable training. We introduce a procedure to certify robustness of 1-Lipschitz CNNs by replacing the last linear layer with a 1-hidden layer. We significantly advance the state-of-the-art for standard and provable robust accuracies on CIFAR-10 and CIFAR-100.
  arXiv  Detail & Related papers  (2022-11-15T19:10:12Z)
• Training \eta-VAE by Aggregating a Learned Gaussian Posterior with a Decoupled Decoder [0.553073476964056]
  Current practices in VAE training often result in a trade-off between the reconstruction fidelity and the continuity$/$disentanglement of the latent space. We present intuitions and a careful analysis of the antagonistic mechanism of the two losses, and propose a simple yet effective two-stage method for training a VAE. We evaluate the method using a medical dataset intended for 3D skull reconstruction and shape completion, and the results indicate promising generative capabilities of the VAE trained using the proposed method.
  arXiv  Detail & Related papers  (2022-09-29T13:49:57Z)
• Mean-Shifted Contrastive Loss for Anomaly Detection [34.97652735163338]
  We propose a new loss function which can overcome failure modes of both center-loss and contrastive-loss methods. Our improvements yield a new anomaly detection approach, based on $textitMean-Shifted Contrastive Loss$. Our method achieves state-of-the-art anomaly detection performance on multiple benchmarks including $97.5%$ ROC-AUC.
  arXiv  Detail & Related papers  (2021-06-07T17:58:03Z)
• Sharp Statistical Guarantees for Adversarially Robust Gaussian Classification [54.22421582955454]
  We provide the first result of the optimal minimax guarantees for the excess risk for adversarially robust classification. Results are stated in terms of the Adversarial Signal-to-Noise Ratio (AdvSNR), which generalizes a similar notion for standard linear classification to the adversarial setting.
  arXiv  Detail & Related papers  (2020-06-29T21:06:52Z)
• The Generalized Lasso with Nonlinear Observations and Generative Priors [63.541900026673055]
  We make the assumption of sub-Gaussian measurements, which is satisfied by a wide range of measurement models. We show that our result can be extended to the uniform recovery guarantee under the assumption of a so-called local embedding property.
  arXiv  Detail & Related papers  (2020-06-22T16:43:35Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.