Coreference Resolution on WSC (Accuracy & Normalized Accuracy)
79.2AccuracyFine-tuned
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Fine-tunedBackbone=T5-large2026.06 | 79.2 | — | |
| Twin-MergingBackbone=T5-large, Merging Strategy=Dynamic model merging2026.06 | 69.7 | 88 | |
| TIES-MergingBackbone=T5-large, Merging Strategy=Static model merging2026.06 | 65.3 | 82.4 | |
| AdaMergingBackbone=T5-large, Merging Strategy=Static model merging2026.06 | 64.4 | 81.3 | |
| DaWinBackbone=T5-large, Merging Strategy=Dynamic model merging2026.06 | 57.9 | 73.1 | |
| WEMoEBackbone=T5-large, Merging Strategy=Dynamic model merging2026.06 | 57 | 72 | |
| Consensus TABackbone=T5-large, Merging Strategy=Static model merging2026.06 | 56 | 70.7 | |
| Task ArithmeticBackbone=T5-large, Merging Strategy=Static model merging2026.06 | 52.8 | 66.7 | |
| MoW-MergingBackbone=T5-large, Merging Strategy=Dynamic model merging2026.06 | 51 | 64.4 | |
| ReTeXBackbone=T5-large, Merging Strategy=Dynamic model merging2026.06 | 48 | 60.6 | |
| Weight AveragingBackbone=T5-large, Merging Strategy=Static model merging2026.06 | 36.1 | 45.6 |