Ranking Consistency Analysis on MMLU-Pro Nutrition health
0.78Spearman CorrelationSCICUSTOM
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| SCICUSTOMConstruction Method=Grounded data and ontology-driven2026.05 | 0.78 | 0.64 | |
| MMLUConstruction Method=Domain-specific comparison2026.05 | 0.56 | 0.51 | |
| GPT-5Construction Method=Synthetic benchmark generation2026.05 | 0.52 | 0.43 | |
| MedQAConstruction Method=Domain-specific comparison2026.05 | 0.45 | 0.31 | |
| EmbeddingConstruction Method=Embedding-based k-NN retrieval2026.05 | 0.27 | 0.18 | |
| GPQAConstruction Method=General scientific benchmark2026.05 | 0.18 | 0.1 | |
| SimpleQAConstruction Method=General-purpose benchmark2026.05 | 0.11 | 0.1 | |
| IfBenchConstruction Method=General-purpose benchmark2026.05 | -0.11 | 0 |