Question Answering on QuaRTz
87.8AccuracyFine-tuned
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Fine-tunedBackbone=T5-large2026.06 | 87.8 | — | |
| Twin-MergingBackbone=T5-large, Merging Strategy=Dynamic model merging2026.06 | 87.1 | 99.2 | |
| ReTeXBackbone=T5-large, Merging Strategy=Dynamic model merging2026.06 | 83.5 | 95.1 | |
| DaWinBackbone=T5-large, Merging Strategy=Dynamic model merging2026.06 | 81.4 | 92.7 | |
| TIES-MergingBackbone=T5-large, Merging Strategy=Static model merging2026.06 | 80.1 | 91.2 | |
| MoW-MergingBackbone=T5-large, Merging Strategy=Dynamic model merging2026.06 | 79 | 90 | |
| Task ArithmeticBackbone=T5-large, Merging Strategy=Static model merging2026.06 | 78.6 | 89.5 | |
| AdaMergingBackbone=T5-large, Merging Strategy=Static model merging2026.06 | 75.5 | 86 | |
| WEMoEBackbone=T5-large, Merging Strategy=Dynamic model merging2026.06 | 74.5 | 84.9 | |
| Weight AveragingBackbone=T5-large, Merging Strategy=Static model merging2026.06 | 70.5 | 80.3 | |
| Consensus TABackbone=T5-large, Merging Strategy=Static model merging2026.06 | 65.5 | 74.6 |