Grokking on RoPE grok depth-8
97.7Post-Grok AccuracyDDC
Evaluation results
| Method | Links | |
|---|---|---|
| DDCSetting=depth-8 RoPE grok, base optimiser=Muon, n=32026.06 | 97.7 | |
| vanilla MuonSetting=depth-8 RoPE grok, base optimiser=Muon, n=32026.06 | 86.7 |
| Method | Links | |
|---|---|---|
| DDCSetting=depth-8 RoPE grok, base optimiser=Muon, n=32026.06 | 97.7 | |
| vanilla MuonSetting=depth-8 RoPE grok, base optimiser=Muon, n=32026.06 | 86.7 |