paper-with-me

홈 › Papers

Topological Signatures of Grokking

2026-05-07 · Yifan Tang, Qiquan Wang, Inés García-Redondo, Anthea Monod arxiv

We study the grokking phenomenon through the lens of topology. Using persistent homology on point clouds derived from the embedding matrices of a range of models trained on modular arithmetic with varying primes, we identify a clear and consistent topological signature of grokking: a sharp increase in both the maximum and total persistence of first homology ($H_1$). Persistence diagrams reveal the emergence of a dominant long-lived topological feature together with increasingly structured secondary features, reflecting the underlying cyclic structure of the task. Compared to existing spectral and geometric diagnostics -- specifically, Fourier analysis and local intrinsic dimension -- persistent homology provides a unified geometric and topological characterization of representation learning, capturing both local and global multi-scale structure. Ablations across data regimes and control settings show that these topological transitions are tied to generalization rather than memorization. Our results suggest that persistent homology offers a principled and interpretable framework for analyzing how neural networks internalize latent structure during training.

📄 PDF Abstract BibTeX arXiv:2605.06352

Code (0)

등록된 구현이 없습니다.

Tasks

Representation LearningPoint Clouds

Similar Papers 제목 키워드 기반

A Stochastic--Geometric Theory of Scaling Laws in Grokking

2026-06-29 · Róisín Luo, Christian Gagné, Jonas Ngnawé, Ihsan Ullah 외 arxiv

Delayed generalization (\ie~grokking) refers to the phenomenon in which a neural network fits its training data early in training but only begins to generalize after a prolonged delay, often through an abrupt transition.…

Topological Signatures of Adversaries in Multimodal Alignments

2025-01-29 · Minh Vu, Geigh Zollicoffer, Huy Mai, Ben Nebgen 외

Multimodal Machine Learning systems, particularly those aligning text and image data like CLIP/BLIP models, have become increasingly prevalent, yet remain susceptible to adversarial attacks. While substantial research ha…

Adversarial Robustness

Deep Learning with Topological Signatures

2017-07-13 · NeurIPS 2017 12 · Christoph Hofer, Roland Kwitt, Marc Niethammer, Andreas Uhl

Inferring topological and geometrical information from data can offer an alternative perspective on machine learning problems. Methods from topological data analysis, e.g., persistent homology, enable us to obtain such i…

BIG-bench Machine LearningDeep LearningTopological Data Analysis

Two Speeds of Learning: A Representation-Readout Decomposition of Grokking and Double Descent

2026-05-26 · Chi-Ning Chou, Oscar Uzdelewicz, Neng-Chun Chiu, Yao-Yuan Yang 외 arxiv

Training loss and accuracy are the standard signals used to monitor generalization during deep neural network training. Two well-documented phenomena complicate this picture: in grokking, train loss falls rapidly while t…

Representation Learning

HalluZig: Hallucination Detection using Zigzag Persistence

2026-01-04 · Shreyas N. Samaga, Gilberto Gonzalez Arroyo, Tamal K. Dey arxiv

The factual reliability of Large Language Models (LLMs) remains a critical barrier to their adoption in high-stakes domains due to their propensity to hallucinate. Current detection methods often rely on surface-level si…