Scaling Law Fitting on LLM-commoncrawl (test)
0.0396Mean Relative ErrorPBH
Evaluation Results
| Method | Links | |
|---|---|---|
| PBHN=5, Dimension (d)=19, LBFGS steps=5, CPU time=300s, Gamma=10, Delta=10^-3, Seeds=32026.05 | 0.0396 | |
| PBHN=15, Dimension (d)=19, LBFGS steps=5, CPU time=300s, Gamma=10, Delta=10^-3, Seeds=32026.05 | 0.0401 | |
| PBHN=10, Dimension (d)=19, LBFGS steps=5, CPU time=300s, Gamma=10, Delta=10^-3, Seeds=32026.05 | 0.0517 | |
| ZOPN=5, Dimension (d)=19, LBFGS steps=5, CPU time=300s, Gamma=10, Delta=10^-3, Seeds=32026.05 | 0.0564 | |
| BHDimension (d)=19, LBFGS steps=5, CPU time=300s, Gamma=10, Delta=10^-3, Seeds=32026.05 | 0.0684 | |
| PBHN=5, lbfgs steps=10, CPU time=300s, d=47, seeds=32026.05 | 0.1045 | |
| PBHN=10, lbfgs steps=10, CPU time=300s, d=47, seeds=32026.05 | 0.1104 | |
| PBHN=50, lbfgs steps=10, CPU time=300s, d=47, seeds=32026.05 | 0.1126 | |
| PBHN=30, lbfgs steps=10, CPU time=300s, d=47, seeds=32026.05 | 0.1145 | |
| BHlbfgs steps=10, CPU time=300s, d=47, seeds=32026.05 | 0.1146 | |
| PBHN=40, lbfgs steps=10, CPU time=300s, d=47, seeds=32026.05 | 0.1152 | |
| ZOPlbfgs steps=10, CPU time=300s, d=47, seeds=32026.05 | 0.1168 | |
| PBHN=20, lbfgs steps=10, CPU time=300s, d=47, seeds=32026.05 | 0.1171 | |
| PBHN=15, lbfgs steps=10, CPU time=300s, d=47, seeds=32026.05 | 0.1183 | |
| PBHN=5, d=47, lbfgs steps=52026.05 | 10.65 | |
| PBHN=20, d=47, lbfgs steps=52026.05 | 11.249 | |
| PBHN=10, d=47, lbfgs steps=52026.05 | 11.254 | |
| PBHN=30, d=47, lbfgs steps=52026.05 | 11.321 | |
| PBHN=50, d=47, lbfgs steps=52026.05 | 11.331 | |
| BHd=47, lbfgs steps=52026.05 | 11.474 | |
| ZOPd=47, lbfgs steps=52026.05 | 11.677 | |
| PBHN=40, d=47, lbfgs steps=52026.05 | 11.679 | |
| PBHN=15, d=47, lbfgs steps=52026.05 | 11.733 |