Two-layer neural networks provably converge almost surely to irreducible representations of finite groups when trained on the group composition task, with the dynamics governed by Riemannian gradient ascent on a representation-theoretic energy functional.
super hub Canonical reference
Grokking: Generalization Beyond Overfitting on Small Algorithmic Datasets
Canonical reference. 94% of citing Pith papers cite this work as background.
101
Pith papers citing it
Background
94% of classified citations
abstract
In this paper we propose to study generalization of neural networks on small algorithmically generated datasets. In this setting, questions about data efficiency, memorization, generalization, and speed of learning can be studied in great detail. In some situations we show that neural networks learn through a process of "grokking" a pattern in the data, improving generalization performance from random chance level to perfect generalization, and that this improvement in generalization can happen well past the point of overfitting. We also study generalization as a function of dataset size and find that smaller datasets require increasing amounts of optimization for generalization. We argue that these datasets provide a fertile ground for studying a poorly understood aspect of deep learning: generalization of overparametrized neural networks beyond memorization of the finite training dataset.
hub tools
citation-role summary
background 16
citation-polarity summary
claims ledger
- abstract In this paper we propose to study generalization of neural networks on small algorithmically generated datasets. In this setting, questions about data efficiency, memorization, generalization, and speed of learning can be studied in great detail. In some situations we show that neural networks learn through a process of "grokking" a pattern in the data, improving generalization performance from random chance level to perfect generalization, and that this improvement in generalization can happen well past the point of overfitting. We also study generalization as a function of dataset size and f
authors
co-cited works
roles
background 16representative citing papers
Derives geodesic ridge regularization and Riemannian Gibbs Process prior for feature-learning wide neural networks, generalizing kernel-regime results via function-space axiomatization.
In the high-dimensional limit the spherical Boltzmann machine admits exact equations for training dynamics, Bayesian evidence, and cascades of phase transitions tied to mode alignment with data, which connect to generative phenomena including double descent and out-of-equilibrium biases.
Transformer weight spectra exhibit transient compression waves that propagate layer-wise, persistent non-monotonic depth gradients in power-law exponents, and Q/K-V asymmetry, with the spectral exponent alpha predicting layer importance and enabling pruning gains of 1.1x-3.6x over Last-N baselines.
Content-based routing succeeds only when models provide bidirectional context and perform pairwise comparisons, with bidirectional Mamba plus rank-1 projection reaching 99.7% precision at linear inference cost.
Infinite-width transformers exhibit an inductive bias against high-complexity polynomial-time algorithms, with derived upper bounds on capturable tasks like sorting and string matching.
Grokking reflects escape from a metastable low-dimensional regime where transverse curvature accumulates before generalization, with subspace motion necessary but curvature boost insufficient.
The AI Scientist framework enables LLMs to independently conduct the full scientific process from idea generation to paper writing and review, demonstrated across three ML subfields with papers costing under $15 each.
Grokking arises from gradual amplification of a Fourier-based circuit in the weights followed by removal of memorizing components.
Toy models demonstrate that polysemanticity arises when neural networks store more sparse features than neurons via superposition, producing a phase transition tied to polytope geometry and increased adversarial vulnerability.
A descent-free method recovers the singularity order k of dead directions in neural networks from the directional-Fisher rate, classifies them, and assembles global learning coefficients matching closed forms.
Observable Matrix Dynamics (OMD) is a new diagnostic framework that uses random matrix theory on distance matrices to distinguish diffusive relaxations from phase-transition-like reorganizations during neural network training.
Dead directions recover Watanabe's RLCT contribution and triple (λ, m, ν) from directional Fisher curvature decay rates in original parameter space for singular models, extended via K-FAC to networks and gauge-equivariant optimizers.
RFLO learning restricts solutions to low-rank perturbations of initial parameters in linear RNNs and produces qualitatively different stability and convergence behavior than BPTT.
Self-evolving rubric with anti-gaming fitness reveals that objective capability scaling fails to transfer to subjective LLM behaviors, with advice-restraint as the universal lowest dimension that can regress.
Apparent phase transitions during fine-tuning on near-synonym tasks are phantoms originating in the softmax readout; an order parameter isolates kinematic and structural failure modes and a few dimensionless quantities predict critical learning rates across architectures via blind test.
Two steps of gradient descent on first-layer weights in linear-width two-layer networks produce a spiked random matrix with floor(alpha2/(1/2-alpha1)) outliers, each a learned direction, and batch reuse allows capturing directions with information exponent exceeding one.
Generalization is a testable hedging property of the learner's response law, recovered via f-divergence regularizers that induce information-geometric curves between training loss and sample dependence.
Temporal correlations from lazy random walks enable efficient SGD learning of k-juntas via temporal-difference loss on ReLU networks, achieving linear sample complexity in d.
Normal alignment is the rank-one Jacobian structure that lets classifiers minimize loss and maximize local robustness in sparse regimes; the paper proves its optimality and uses it to create GrokAlign and RFAMs.
Persistent homology detects a sharp increase in maximum and total H1 persistence during grokking on modular arithmetic, offering a topological diagnostic that links representation geometry to generalization.
Slingshot loss spikes are produced by low-precision arithmetic that breaks the zero-sum gradient constraint and drives exponential growth via Numerical Feature Inflation.
Gradient matching empirically recovers implicit regularization effects such as l2 penalties from early stopping and dropout in neural networks.
Steepest descent under divergence-induced quadratic models equals an LQR problem, enabling learning of diagonal or Kronecker-factored inverse preconditioners via a global layerwise objective for scalable geometry-aware training.
citing papers explorer
-
Transformers Linearly Represent Highly Structured World Models
Transformers trained on Sudoku traces develop constraint-structured internal world models and a monosemantic naked-single circuit.
-
Detecting overfitting in Neural Networks during long-horizon grokking using Random Matrix Theory
Random Matrix Theory detects overfitting via growing Correlation Traps in weight spectra during the anti-grokking phase of neural network training.
-
Overtrained, Not Misaligned
Emergent misalignment arises from overtraining after primary task convergence and is preventable by early stopping, which retains 93% of task performance on average.
-
The two clocks and the innovation window: When and how generative models learn rules
Generative models learn rules before memorizing data, creating an innovation window whose width depends on dataset size and rule complexity, observed in both diffusion and autoregressive architectures.
-
Selection Plateau and a Sparsity-Dependent Hierarchy of Pruning Features
All rank-monotone pruning scorers converge to identical accuracy at fixed sparsity, but non-monotone features with sparsity-dependent complexity can escape this plateau, as shown by the SICS hypothesis on ViT-Small/CIFAR-10.
-
Learning Large-Scale Modular Addition with an Auxiliary Modulus
An auxiliary modulus during training reduces wrap-around issues and preserves train-test input distributions, enabling better accuracy and sample efficiency for large N and q in modular addition learning.
-
The Weight Gram Matrix Captures Sequential Feature Linearization in Deep Networks
Gradient descent in deep networks implicitly drives features toward target-linear structure as captured by the weight Gram matrix and a derived virtual covariance.
-
Spectral Lens: Activation and Gradient Spectra as Diagnostics of LLM Optimization
Spectral analysis of activations and gradients provides new diagnostics that link batch size to representation geometry, early covariance tails to token efficiency, and spectral shifts to learning dynamics in decoder-only LLMs, backed by a mechanistic model.
-
Critical Windows of Complexity Control: When Transformers Decide to Reason or Memorize
Transformers show a sharp, task-specific critical window for weight decay application that determines reasoning versus memorization, with middle placement optimal and boundaries as narrow as 100 steps.
-
Can Transformers predict system collapse in dynamical systems?
Transformers fail to predict catastrophic collapse in unseen parameter regimes of nonlinear dynamical systems, while reservoir computing reliably succeeds.
-
Finite-Size Gradient Transport in Large Language Model Pretraining: From Cascade Size to Intensive Transport Efficiency
A gradient-transport framework with observables D, z, β, δ, v_rel applied to Pico-LM and Pythia datasets shows distinct scaling regimes in duration and efficiency while sharing a near-unity cascade-size backbone.
-
Convergent Evolution: How Different Language Models Learn Similar Number Representations
Diverse language models converge on similar periodic number features with a two-tier hierarchy of Fourier sparsity and geometric separability, acquired via language co-occurrences or multi-token arithmetic.
-
Generalization at the Edge of Stability
Training at the edge of stability causes neural network optimizers to converge on fractal attractors whose effective dimension, measured via a new sharpness dimension from the Hessian spectrum, bounds generalization error in a way not captured by prior trace or norm measures.
-
Spectral Entropy Collapse as a Phase Transition in Delayed Generalisation: An Interventional and Predictive Framework for Grokkin
Spectral entropy collapse in learned representations precedes and predicts grokking, with interventions showing it is not explained by parameter norm alone.
-
Training Deep Visual Networks Beyond Loss and Accuracy Through a Dynamical Systems Approach
Introduces integration, metastability, and dynamical stability index measures from layer activations and reports patterns distinguishing CIFAR-10 from CIFAR-100 difficulty plus early convergence signals across ResNet variants, DenseNet, MobileNetV2, VGG-16, and a Vision Transformer.
-
Grokking as Dimensional Phase Transition in Neural Networks
Grokking occurs as the effective dimensionality of the gradient field transitions from sub-diffusive to super-diffusive at the onset of generalization, exhibiting self-organized criticality.
-
Autolearn: Learn by Surprise, Commit by Proof
Autolearn uses high-loss passages and self-generated Q&A training to drive the perturbation gap below baseline, improving novel fact acquisition while suppressing memorization in language models.
-
Spectral Edge Dynamics: An Analytical-Empirical Study of Phase Transitions in Neural Network Training
Spectral gaps in the Gram matrix of parameter updates control phase transitions such as grokking in neural network training.
-
AInstein: Can LLMs Solve Research Problems From Parametric Memory Alone?
LLMs generate valid solutions to over 70% of AI research problems from parametric memory alone but rediscover the exact published approach less than 19% of the time, with performance limited by cross-domain analogical transfer.
-
Feature Identification via the Empirical NTK
Eigenanalysis of the empirical NTK surfaces feature directions that align with Fourier features in modular addition networks and grammatical features in Gemma-3-270M, outperforming PCA baselines on activations.
-
Emergent Slow Thinking in LLMs as Inverse Tree Freezing
RLVR drives a concept network in LLMs through nucleation and freezing into inverse trees that support slow thinking, and intervening with brief SFT at peak frustration outperforms standard RLVR while post-freeze SFT causes forgetting.
-
ShinkaEvolve: Towards Open-Ended And Sample-Efficient Program Evolution
ShinkaEvolve improves sample efficiency in LLM-driven program evolution via parent sampling, code novelty rejection-sampling, and bandit LLM ensemble selection, achieving new SOTA circle packing with 150 samples and gains on math reasoning and competitive programming tasks.
-
MetaMorph: Multimodal Understanding and Generation via Instruction Tuning
VPiT enables pretrained LLMs to perform both visual understanding and generation by predicting discrete text tokens and continuous visual tokens, with understanding data proving more effective than generation-specific data.
-
Language Models (Mostly) Know What They Know
Language models show good calibration when asked to estimate the probability that their own answers are correct, with performance improving as models get larger.
-
Two AI Metrics Diverged: Will it Make All the Difference?
Bounded performance metrics always favor convergence of AI capabilities to meek models while unbounded metrics allow frontier models to maintain leads indefinitely, with policy implications for capability concentration.
-
World-Model Collapse as a Phase Transition
Long-horizon language agents show phase-transition-like world-model collapse under small parameter changes, with world-state fidelity failing before action validity, as mapped by grid search in deterministic tasks with gold states.
-
Scalar Representations of Neural Network Training Dynamics
Scalar embeddings of neural network training trajectories treated as temporal networks preserve main dynamical features including Lyapunov exponents, enable definition of a characteristic decorrelation time, and show asymptotic state spacings compatible with a skew lognormal distribution.
-
SingGuard: A Policy-Adaptive Multimodal LLM Guardrail with Dynamic Reasoning
SingGuard introduces a policy-adaptive multimodal LLM guardrail with dynamic reasoning regimes and SingGuard-Bench, reporting SOTA F1 scores across 35 datasets and improved policy-following accuracy under runtime shifts.
-
Structure and Scale in Simplicial Sequence Modelling
Small transformers on HMM prediction tasks exhibit correlated scaling between performance and linear encoding of belief distributions in residual activations.
-
Comparing Classical Simulation and Sample-Based Learning of Quantum Systems: Learning the Hardness of Quantum Systems from Samples
Empirical study finds neural-network learning difficulty (via Hessian eigenvalue and random subspace optimization) correlates with classical simulation hardness parameterized by MPS bond dimension and T-gate count.
-
The Future of Facts: Tracing the Factual Generation-Verification Gap
Empirical tracing across model families shows verification precedes and outlasts generation for facts, with updates producing simultaneous verification of old and new answers.
-
Emergence via Phase Transitions: Mechanism Landscapes and Universal Convergence Across Complex Systems
HEF frames emergence as a phase transition in mechanism landscapes with proofs of convergence under structural assumptions, empirically tested via grokking in transformers showing weight-norm peaks and accuracy collapse to ~0.9745.
-
Axiomatizing Neural Networks via Pursuit of Subspaces
Authors introduce the Pursuit of Subspaces (PoS) hypothesis, an axiomatic geometric framework that unifies explanations for representation, computation, and generalization in shallow and deep neural networks.
-
Unveiling Memorization-Generalization Coexistence: A Case Study on Arithmetic Tasks with Label Noise
Experiments on modular arithmetic with heavy label noise show that over-parameterized networks form a distributed internal generalization structure that can be extracted via frequency methods to achieve high accuracy despite 80% noise.
-
Constrained Stochastic Spectral Preconditioning Converges for Nonconvex Objectives
Proximal stochastic spectral preconditioning converges for nonconvex constrained objectives under heavy-tailed noise, with a variance-reduced version achieving faster rates and a refined analysis of Muon iterations.
-
Model Capacity Determines Grokking through Competing Memorisation and Generalisation Speeds
Grokking emerges near the model size where memorization timescale T_mem(P) intersects generalization timescale T_gen(P) on modular arithmetic.
-
Phase Transitions in Driven Informational Systems: A Two-Field Perspective on Learning Theory and Non-Equilibrium Chemistry
Proposes a two-gradient-field model with candidate order parameters alpha_dagger and kappa_c to unify phase transitions across learning theory and non-equilibrium chemistry.
-
Universal Quantum Transformer
UQT on 5 qubits achieves exact deterministic learning of Z_11 modular arithmetic and S_4 non-Abelian algebra via quantum-native mechanisms, claiming to bypass classical attention limits and run on NISQ hardware.
-
Gradient-Direction Sensitivity Reveals Linear-Centroid Coupling Hidden by Optimizer Trajectories
Gradient-based SVD diagnostic uncovers hidden SED-LCH coupling in single and multitask settings and shows rank-3 subspace constraints speed up grokking by 2.3x.
-
On the Convergence Behavior of Preconditioned Gradient Descent Toward the Rich Learning Regime
Preconditioned gradient descent mitigates spectral bias and reduces grokking delays by enabling uniform parameter space exploration in the NTK regime, confirming grokking as a transition to the rich regime.
-
FunduSegmenter: Leveraging the RETFound Foundation Model for Joint Optic Disc and Optic Cup Segmentation in Retinal Fundus Images
FunduSegmenter adapts RETFound with custom adapters and attention modules to reach 90.51% average Dice score for optic disc and cup segmentation, outperforming baselines on internal and external tests across five datasets.
-
Learning the symmetric group: large from small
Transformer trained on S10 permutation prediction from transpositions generalizes to S25 with near 100% accuracy using identity augmentation and partitioned windows.
-
Unified Neural Scaling Laws
Presents a single functional form for neural scaling that unifies multiple scaling dimensions and claims higher extrapolation accuracy than prior forms across diverse tasks and architectures.
-
Artificial Jagged Intelligence as Uneven Optimization Energy Allocation Capability Concentration, Redistribution, and Optimization Governance
AJI frames jagged AI capabilities as lower bounds on performance dispersion arising from concentrated optimization energy allocation under anisotropic objectives, with theorems on tradeoffs and redistribution interventions.
-
Feature Repulsion and Spectral Lock-in: An Empirical Study of Two-Layer Network Grokking
Empirical tests confirm robust feature repulsion signs but reveal activation-dependent spectral lock-in in grokking, with x^2 yielding rank-2 updates at epoch ~174 and ReLU remaining rank-1.
-
Step-Video-T2V Technical Report: The Practice, Challenges, and Future of Video Foundation Model
Step-Video-T2V describes a 30B-parameter text-to-video model with custom Video-VAE, 3D DiT, flow matching, and Video-DPO that claims state-of-the-art results on a new internal benchmark.
-
A Survey of Large Language Models
This survey reviews the background, key techniques, and evaluation methods for large language models, emphasizing emergent abilities that appear at large scales.
-
When AI meets quantum information: A comprehensive review
A comprehensive review organizing progress at the AI-quantum information intersection from both directions.
- Nexus: Same Pretraining Loss, Better Downstream Generalization via Common Minima
- The Long Delay to Arithmetic Generalization: When Learned Representations Outrun Behavior