[2603.29262] Grokking From Abstraction to Intelligence

Computer Science > Artificial Intelligence arXiv:2603.29262 (cs) [Submitted on 31 Mar 2026] Title:Grokking From Abstraction to Intelligence Authors:Junjie Zhang, Zhen Shen, Gang Xiong, Xisong Dong View a PDF of the paper titled Grokking From Abstraction to Intelligence, by Junjie Zhang and 3 other authors View PDF HTML (experimental) Abstract:Grokking in modular arithmetic has established itself as the quintessential fruit fly experiment, serving as a critical domain for investigating the mechanistic origins of model generalization. Despite its significance, existing research remains narrowly focused on specific local circuits or optimization tuning, largely overlooking the global structural evolution that fundamentally drives this phenomenon. We propose that grokking originates from a spontaneous simplification of internal model structures governed by the principle of parsimony. We integrate causal, spectral, and algorithmic complexity measures alongside Singular Learning Theory to reveal that the transition from memorization to generalization corresponds to the physical collapse of redundant manifolds and deep information compression, offering a novel perspective for understanding the mechanisms of model overfitting and generalization. Comments: Subjects: Artificial Intelligence (cs.AI) Cite as: arXiv:2603.29262 [cs.AI]   (or arXiv:2603.29262v1 [cs.AI] for this version)   https://doi.org/10.48550/arXiv.2603.29262 Focus to learn more arXiv-issued DOI via DataCite (pending...