New paper detects overfitting directly from weight matrices without any data
burny_tech · x · 2026-09-20
- The WeightWatcher paper 'Late-Stage Generalization Collapse in Grokking: Detecting Anti-Grokking with WeightWatcher' shows you can detect overfitting signatures directly in layer weight matrices, without access to any train or test data.
- Demonstrated explicitly in two classic grokking experiments, revealing a form of late-stage overfitting different from what has been seen before.
- The authors track the Grokking → Stability → Anti-Grokking trajectory and have submitted the work to ICLR.
Related event: Paper Reveals Third Stage of Grokking: Late-Stage Generalization Collapse(2 posts)→
More from Research
- Anthropic says it may imminently hit RSP threshold for AI-accelerated R&D, releases provable inference work — tobyordoxford · 2026-11-04
- Google scanned 1,033 real household products, packed into a 345 MB GLB for three.js — victormustar · 2026-11-04
- These 10 GPU optimization papers are why your model runs fast today — NandoDF · 2026-11-04
- INT8 isn't the problem: I benchmarked samplers on INT8 flow models — YakovSava · 2026-11-04
- Karpathy Boosts Open Hacker News Dataset: 10M+ Posts Usable as a Personal Knowledge Base — huangyun_122 · 2026-11-04
- PixelCraft: A Multi-Agent Framework Targets Multimodal AI's Chart-Reading Failures — burkov · 2026-11-04