Multi-Hop Question Answering on Bamb (F1 score)
71F1 Score (Bamb)ReCal
Evaluation Results
| Method | Links | |
|---|---|---|
| ReCalRouting=RL-based Routing, Variant=+2026.06 | 71 | |
| ReCalRouting=RL-based Routing2026.06 | 69 | |
| RouterDCRouting=Heuristic & Discriminative Routing2026.06 | 62.3 | |
| Largest LLMRouting=Heuristic & Discriminative Routing2026.06 | 60.8 | |
| Router-R1Routing=RL-based Routing2026.06 | 59.6 | |
| Prompt LLM*Routing=Heuristic & Discriminative Routing2026.06 | 58 | |
| Prompt LLMRouting=Heuristic & Discriminative Routing2026.06 | 57.9 | |
| GraphRouterRouting=Heuristic & Discriminative Routing2026.06 | 54.6 | |
| KNN RouterRouting=Heuristic & Discriminative Routing2026.06 | 49.6 | |
| KNN Router*Routing=Heuristic & Discriminative Routing2026.06 | 49.3 | |
| MLP RouterRouting=Heuristic & Discriminative Routing2026.06 | 46 | |
| RouteLLMRouting=Heuristic & Discriminative Routing2026.06 | 42.3 | |
| Search-R1Routing=No Routing2026.06 | 38.7 | |
| CoTRouting=No Routing2026.06 | 33.2 | |
| RAGRouting=No Routing2026.06 | 30.3 | |
| SFTRouting=No Routing2026.06 | 17.3 | |
| VanillaRouting=No Routing2026.06 | 11.2 |