RAG-Completeness on ALCE (test)
0.47Mean Kendall's TauJury-on-Demand
Evaluation Results
| Method | Links | |
|---|---|---|
| Jury-on-DemandJury Strategy=Dynamic2025.12 | 0.47 | |
| GPT-OSS-20BJury Strategy=Single Judge2025.12 | 0.4 | |
| Static (Avg-All)Jury Strategy=Static, Aggregation=Average All2025.12 | 0.38 | |
| Static (W-Reg)Jury Strategy=Static, Aggregation=Weighted-Regression2025.12 | 0.34 | |
| Static (Avg-TopK)Jury Strategy=Static, Aggregation=Average TopK2025.12 | 0.28 | |
| Static (W-Tau)Jury Strategy=Static, Aggregation=Weighted-Tau2025.12 | 0.23 |