Cover Image

The Geometric Inductive Bias of Grokking: Bypassing Phase Transitions via Architectural Topology

Fuente: arXiv

Saved in:

Bibliographic Details
Main Author:	Yıldırım, Alper
Format:	Preprint
Published:	2026
Subjects:	Machine Learning Artificial Intelligence I.2.6; F.2
Online Access:	Acceder al recurso
Tags:	Add Tag No Tags, Be the first to tag this record!

Similar Items

Superposition Is Not Necessary: A Mechanistic Interpretability Analysis of Transformer Representations for Time Series Forecasting
by: Yıldırım, Alper
Published: (2026)

Language as a Wave Phenomenon: Semantic Phase Locking and Interference in Neural Networks
by: Yıldırım, Alper, et al.
Published: (2025)

Grokking Beyond the Euclidean Norm of Model Parameters
by: Notsawo, Pascal Jr Tikeng, et al.
Published: (2025)

Interestingness as an Inductive Heuristic for Future Compression Progress
by: Herrmann, Vincent, et al.
Published: (2026)

Hidden Failure Modes of Gradient Modification under Adam in Continual Learning, and Adaptive Decoupled Moment Routing as a Repair
by: Hu, Yuelin, et al.
Published: (2026)

Intervention-Assisted Policy Gradient Methods for Online Stochastic Queuing Network Optimization: Technical Report
by: Wigmore, Jerrod, et al.
Published: (2024)

The Architecture of Errors: From Universal Impossibility to Patch-Local LLM Reliability
by: Arbuzov, Mikhail L., et al.
Published: (2026)

Deep Memory Search: A Metaheuristic Approach for Optimizing Heuristic Search
by: Hedar, Abdel-Rahman, et al.
Published: (2024)

Time-to-Injury Forecasting in Elite Female Football: A DeepHit Survival Approach
by: Catterall, Victoria, et al.
Published: (2026)

Reducing the Sensitivity of Neural Physics Simulators to Mesh Topology via Pretraining
by: Vaska, Nathan, et al.
Published: (2025)

A Geometric Perspective for High-Dimensional Multiplex Graphs
by: Abdous, Kamel, et al.
Published: (2025)

Not All Transitions Matter: Evidence from PPO
by: Basnet, Ajhesh
Published: (2026)

Quantization Undoes Alignment: Bias Emergence in Compressed LLMs Across Models and Precision Levels
by: Rath, Plawan Kumar, et al.
Published: (2026)

Grokking in the Wild: Data Augmentation for Real-World Multi-Hop Reasoning with Transformers
by: Abramov, Roman, et al.
Published: (2025)

Differentiable Symbolic Planning: A Neural Architecture for Constraint Reasoning with Learned Feasibility
by: Oruganti, Venkatakrishna Reddy
Published: (2026)

Architectural Proprioception in State Space Models: Thermodynamic Training Induces Anticipatory Halt Detection
by: Noon, Jay
Published: (2026)

Leveraging Personalized PageRank and Higher-Order Topological Structures for Heterophily Mitigation in Graph Neural Networks
by: Wang, Yumeng, et al.
Published: (2025)

TRIM: Achieving Extreme Sparsity with Targeted Row-wise Iterative Metric-driven Pruning
by: Beck, Florentin, et al.
Published: (2025)

Model Fusion via Retrofitting
by: Luenam, Phoomraphee, et al.
Published: (2025)

Learning Nonlinear Regime Transitions via Semi-Parametric State-Space Models
by: Hiremath, Prakul Sunil
Published: (2026)

BandiK: Efficient Multi-Task Decomposition Using a Multi-Bandit Framework
by: Millinghoffer, András, et al.
Published: (2025)

Semi-overlapping Multi-bandit Best Arm Identification for Sequential Support Network Learning
by: Antos, András, et al.
Published: (2025)

Democratic Preference Alignment via Sortition-Weighted RLHF
by: Sana, Suvadip, et al.
Published: (2026)

BLOCK-EM: Preventing Emergent Misalignment via Latent Blocking
by: Ustaomeroglu, Muhammed, et al.
Published: (2026)

Scaling Offline RL via Efficient and Expressive Shortcut Models
by: Espinosa-Dice, Nicolas, et al.
Published: (2025)

Neuro-Symbolic Learning for Galois Groups: Unveiling Probabilistic Trends in Polynomials
by: Shaska, Elira, et al.
Published: (2025)

Embedded Safety-Aligned Intelligence via Differentiable Internal Alignment Embeddings
by: Rathva, Harsh, et al.
Published: (2025)

GIRL: Generative Imagination Reinforcement Learning via Information-Theoretic Hallucination Control
by: Hiremath, Prakul Sunil
Published: (2026)

CircuitProbe: Predicting Reasoning Circuits in Transformers via Stability Zone Detection
by: Panuganti, Rajkiran
Published: (2026)

FastForward Pruning: Efficient LLM Pruning via Single-Step Reinforcement Learning
by: Yuan, Xin, et al.
Published: (2025)

FlowSteer: Towards Agents Designing Agentic Workflows via Reinforced Progressive Canvas Editing
by: Zhang, Mingda, et al.
Published: (2026)

Leveraging KANs for Expedient Training of Multichannel MLPs via Preconditioning and Geometric Refinement
by: Actor, Jonas A., et al.
Published: (2025)

FastGRPO: Accelerating Policy Optimization via Concurrency-aware Speculative Decoding and Online Draft Learning
by: Zhang, Yizhou, et al.
Published: (2025)

SaiVLA-0: Cerebrum--Pons--Cerebellum Tripartite Architecture for Compute-Aware Vision-Language-Action
by: Shi, Xiang, et al.
Published: (2026)

ReflexGrad: Within-Episode Failure Recovery in LLM Agents via Progress-Gated Dual-Process Routing
by: Kadu, Ankush, et al.
Published: (2025)

Bias by Necessity: Impossibility Theorems for Sequential Processing with Convergent AI and Human Validation
by: Wu, Jikun, et al.
Published: (2026)

DELTA: Variational Disentangled Learning for Privacy-Preserving Data Reprogramming
by: Malarkkan, Arun Vignesh, et al.
Published: (2025)

Spectral Compact Training: Pre-Training Large Language Models via Permanent Truncated SVD and Stiefel QR Retraction
by: Kohlberger, Björn Roman
Published: (2026)

First-Mover Bias in Gradient Boosting Explanations: Mechanism, Detection, and Resolution
by: Caraker, Drake, et al.
Published: (2026)

On the Push-Based Asynchronous Federated Learning: A Bias-Correction Aggregation Approach
by: Bai, Jiahui, et al.
Published: (2026)