A residual from Hankel DMD on Wasserstein-mapped training distributions localizes grokking transitions in modular-addition Transformers with AUROC 0.93 and can precede onset under a sustained-threshold rule.
The result quantifies sensitivity, not causal mechanism; one boundary-case low-RR early-training failure is reported alongside the main pool
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Distributional Spectral Diagnostics for Localizing Grokking Transitions
A residual from Hankel DMD on Wasserstein-mapped training distributions localizes grokking transitions in modular-addition Transformers with AUROC 0.93 and can precede onset under a sustained-threshold rule.