Fact Memorization on Wikipedia corpus annotated (train)
929.25Fact AccuracyLossH-Wiki
Evaluation results
| Method | Links | |
|---|---|---|
| LossH-Wikialpha (α)=0.2, Model size=110m2026.04 | 929.25 | |
| LossH-Wikialpha (α)=0.3, Model size=110m2026.04 | 914.87 | |
| LossH-Wikialpha (α)=0.4, Model size=110m2026.04 | 880.15 | |
| LossH-Wikialpha (α)=0.5, Model size=110m2026.04 | 851.69 | |
| LossH-Wikialpha (α)=0.1, Model size=110m2026.04 | 835.81 | |
| LossH-Wikialpha (α)=0.6, Model size=110m2026.04 | 829.15 | |
| LossH-Wikialpha (α)=0.7, Model size=110m2026.04 | 790.76 | |
| LossH-Wikialpha (α)=0.8, Model size=110m2026.04 | 763.6 | |
| LossH-Wikialpha (α)=0.9, Model size=110m2026.04 | 733.23 | |
| Random Baselinealpha (α)=1.0, Model size=110m2026.04 | 696.11 | |
| LossH-Wikialpha (α)=1.0, Model size=110m2026.04 | 696.11 | |
| Random Baselinealpha (α)=0.9, Model size=110m2026.04 | 690.86 | |
| Random Baselinealpha (α)=0.8, Model size=110m2026.04 | 687.72 | |
| Random Baselinealpha (α)=0.6, Model size=110m2026.04 | 687.23 | |
| Random Baselinealpha (α)=0.7, Model size=110m2026.04 | 685.46 | |
| Random Baselinealpha (α)=0.5, Model size=110m2026.04 | 676.29 | |
| Random Baselinealpha (α)=0.4, Model size=110m2026.04 | 666.84 | |
| Random Baselinealpha (α)=0.3, Model size=110m2026.04 | 657.25 | |
| Random Baselinealpha (α)=0.2, Model size=110m2026.04 | 647.33 | |
| Random Baselinealpha (α)=0.1, Model size=110m2026.04 | 600.55 |