On MNIST with a three-layer ReLU MLP and AdamW, every EGD variant reaches 85% validation accuracy in fewer steps than Grokfast. RSVD variants also substantially reduce wall-clock time; rank 4 is fastest in time, ranks 8 and 4 tie for the fewest steps, and exact-SVD EGD has the highest printed final validation accuracy.

On MNIST with a three-layer ReLU MLP and AdamW, every EGD variant reaches 85% validation accuracy in fewer steps than Grokfast. RSVD variants also substantially reduce wall-clock time; rank 4 is fastest in time, ranks 8 and 4 tie for the fewest steps, and exact-SVD EGD has the highest printed final validation accuracy. · CiteArk

Measurement results

0/34 assessed

0 supported0 disputed or conflicting34 not assessed

The complete result group stays with its claim. Assessed rows come first; their judgments do not replace the whole-claim conclusion. Expand a row for conditions, sources, and evidence history.

Each row is a measurement with its reported value, latest available observation, and cumulative evidence state.
Metric and conditions	Reported	Latest observation	Evidence judgment
method: Vanilla AdamW	43,100 optimization steps	—	Not assessed
method: Vanilla AdamW	660 seconds	—	Not assessed
method: Vanilla AdamW	15.3 milliseconds	—	Not assessed
method: Vanilla AdamW	89%	—	Not assessed
method: Grokfast	1,800 optimization steps	—	Not assessed
method: Grokfast	23.9 fold	—	Not assessed
method: Grokfast	28.7 seconds	—	Not assessed
method: Grokfast	23 fold	—	Not assessed
method: Grokfast	16 milliseconds	—	Not assessed
method: Grokfast	91%	—	Not assessed
method: EGD (SVD)	700 optimization steps	—	Not assessed
method: EGD (SVD)	61.6 fold	—	Not assessed

Reported and observed measurements

Assessments (0)

Runs (0)

Measurement results