Multi-task Language Understanding on MMLU (in-distribution)
74.4AccuracyCONCUR
Evaluation results
| Method | Links | ||
|---|---|---|---|
| CONCURRouting=Unconstrained2025.12 | 74.4 | 40.73 | |
| Best single strategyRouting=Unconstrained2025.12 | 73.7 | 44.45 | |
| RTRRouting=Unconstrained2025.12 | 73.7 | 44.45 | |
| EmbedLLMRouting=Unconstrained2025.12 | 72.1 | 33.68 | |
| RouteLLMRouting=Unconstrained2025.12 | 54.7 | 3.2 |