SOURCE-LINKED INTELLIGENCE

Beyond Embedding Transfer: Component Roles in Grokking Transfer and Stability

arXiv · AI, language, vision and robotics · article · Sep 16, 2026 · UTC

Warm-start transfer can make algorithmic tasks generalize rapidly, yet it is unclear which model components provide the gain and whether that gain remains stable under continued optimization. We study cross-operator transfer on modular arithmetic and separate efficacy (early velocity) from stability (post-reach drawdown). In a scale-matched 108-run battery across 12 seed blocks (96-run 2^3 factorial plus 12-run scale control), transferring internal attention/MLP weights (B) alongside token embeddings and readout (E+U) improves early accuracy by 5.46 pp (Holm p=0.0039) and cuts confirmation lat

Read original source ↗ Open in workspace

recordType

paper

region

Global