Curator Triage on BioConCal Document-level (test)
95.2Test PrecisionUnanimous / raw agreement P95
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| Unanimous / raw agreement P95Cand. source=LLM-82026.05 | 95.2 | 279 | 14 | 293 | 11.6 | |
| BioConCal GBTCand. source=LLM-82026.05 | 93.9 | 1,258 | 82 | 1,340 | 52.2 | |
| BioConCal-HybridCand. source=merged union2026.05 | 92.9 | 1,716 | 131 | 1,847 | 70.9 | |
| LLM-8 ∪ BERN2 exploratory mergeCand. source=additive union2026.05 | 83.6 | 1,323 | 260 | 1,583 | 55 | |
| BERN2 take-allCand. source=BERN22026.05 | 80.6 | 1,772 | 427 | 2,199 | 73.7 | |
| Agreement ≥ 2Cand. source=LLM-82026.05 | 69.2 | 1,790 | 797 | 2,587 | 74.5 |