{
  "id": 9256663,
  "title": "A Spectral Theory of Grokking: Weight Decay induces Feature Learning",
  "url": "https://urgent.news/2026/09/22/a-spectral-theory-of-grokking-weight-decay-induces-feature-learning",
  "topic": "ai",
  "section": "AI",
  "published": "2026-09-22T16:38:16.000Z",
  "source": {
    "name": "arXiv cs.AI",
    "slug": "arxiv-cs-ai",
    "url": "https://arxiv.org/abs/2609.26679v1"
  },
  "original_language": "en",
  "account": null,
  "summary": "In grokking an early fit to the training data separates from a much later improvement in generalization. During this delay, training can move from a fixed neural tangent kernel (NTK) regime to one in which task-relevant kernel eigendirections continue to evolve. We provide a quantitative theory for how this transition from lazy to rich learning can produce delayed generalization. For homogeneous…",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}