For modular multiplication at p=97 with a two-layer transformer and Adam, Grokfast reaches 95% accuracy sooner than exact-SVD EGD in both steps and wall-clock time. RSVD rank 36 is closer to Grokfast but remains slower by the printed values; all three accelerated methods finish at 100% validation accuracy, versus 99.1% for vanilla Adam.

For modular multiplication at p=97 with a two-layer transformer and Adam, Grokfast reaches 95% accuracy sooner than exact-SVD EGD in both steps and wall-clock time. RSVD rank 36 is closer to Grokfast but remains slower by the printed values; all three accelerated methods finish at 100% validation accuracy, versus 99.1% for vanilla Adam. · CiteArk

Measurement results

0/22 assessed

0 supported0 disputed or conflicting22 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 Adam	55,780 optimization steps	—	Not assessed
method: Vanilla Adam	5,198.7 seconds	—	Not assessed
method: Vanilla Adam	93.2 milliseconds	—	Not assessed
method: Vanilla Adam	99.1%	—	Not assessed
method: Grokfast	910 optimization steps	—	Not assessed
method: Grokfast	61.3 fold	—	Not assessed
method: Grokfast	114 seconds	—	Not assessed
method: Grokfast	45.6 fold	—	Not assessed
method: Grokfast	125.3 milliseconds	—	Not assessed
method: Grokfast	100%	—	Not assessed
method: EGD (SVD)	1,170 optimization steps	—	Not assessed
method: EGD (SVD)	47.7 fold	—	Not assessed

Reported and observed measurements

Assessments (0)

Runs (0)

Measurement results