Large Model Performance Prediction Dataset 80% masking (test)
7.5RMSESTAR
Evaluation results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| STARMasking Ratio=80%2026.02 | 7.5 | 4.92 | 6.21 | 97.04 | 90.18 | 51.31 | 79.18 | — | |
| Semantic-Augmented CPMFMasking Ratio=80%2026.02 | 8.75 | 5.36 | 7.06 | 96.5 | 89.64 | 50.28 | 78.81 | — | |
| PMFMasking Ratio=80%, Category=Statistic-based Methods2026.02 | 10.56 | 6.81 | 8.69 | 95.02 | 86.87 | 46.52 | 76.17 | — | |
| NCFMasking Ratio=80%, Category=Statistic-based Methods2026.02 | 11.51 | 7.79 | 9.65 | 92.68 | 84.19 | 28.64 | 68.5 | — | |
| RegressionMasking Ratio=80%, Category=Statistic-based Methods2026.02 | 11.98 | 8.19 | 10.09 | 92.72 | 84.78 | 44.86 | 74.12 | — | |
| Mean Of MeansMasking Ratio=80%, Category=Statistical Baseline2026.02 | 12.88 | 9.69 | 11.29 | 92.73 | 84.56 | 47.36 | 74.89 | 63.6 | |
| PCAMasking Ratio=80%, Category=Statistic-based Methods2026.02 | 14.54 | 10.81 | 12.67 | 88.03 | 79.26 | 22.19 | 63.16 | — | |
| LLM-RAGMasking Ratio=80%, Category=LLM-based Methods2026.02 | 14.72 | 8.98 | 11.85 | 77.5 | 62.29 | 29.35 | 56.38 | — | |
| Global MeanMasking Ratio=80%, Category=Statistical Baseline2026.02 | 16.18 | 12.94 | 14.56 | 85.56 | 76.95 | 12.12 | 58.21 | 43.65 | |
| Pure LLMMasking Ratio=80%, Category=LLM-based Methods2026.02 | 23.5 | 16.63 | 20.07 | 72.03 | 66.26 | 24.27 | 54.25 | — |