Model Merging on Average of 8 benchmarks
52.79Average AccuracyPico
Evaluation Results
| Method | Links | |
|---|---|---|
| PicoRank=32, Merge Method=Task Arithmetic2026.04 | 52.79 | |
| PicoRank=64, Merge Method=Task Arithmetic2026.04 | 52.21 | |
| PicoRank=32, Merge Method=TIES2026.04 | 50.43 | |
| KnOTSRank=32, Merge Method=TIES2026.04 | 50.29 | |
| PicoRank=8, Merge Method=TIES2026.04 | 50.08 | |
| TIESRank=32, Calibration=No Calibration2026.04 | 50.05 | |
| KnOTSRank=8, Merge Method=TIES2026.04 | 49.97 | |
| PicoRank=64, Merge Method=TIES2026.04 | 49.41 | |
| PicoRank=8, Merge Method=Task Arithmetic2026.04 | 48.9 | |
| DARERank=64, Merge Method=TIES2026.04 | 48.78 | |
| KnOTSRank=64, Merge Method=TIES2026.04 | 48.7 | |
| TIESRank=64, Calibration=No Calibration2026.04 | 48.58 | |
| DARERank=8, Merge Method=TIES2026.04 | 48.57 | |
| PicoRank=8, Merge Method=TSV-M2026.04 | 48.47 | |
| DARERank=32, Merge Method=TIES2026.04 | 48.27 | |
| KnOTSRank=32, Merge Method=Task Arithmetic2026.04 | 48.23 | |
| TIESRank=8, Calibration=No Calibration2026.04 | 48.16 | |
| Core SpaceRank=32, Merge Method=Task Arithmetic2026.04 | 48.02 | |
| Core SpaceRank=32, Merge Method=TIES2026.04 | 48.01 | |
| Task ArithmeticRank=32, Calibration=No Calibration2026.04 | 47.54 | |
| Core SpaceRank=64, Merge Method=TIES2026.04 | 47.49 | |
| DELLARank=32, Merge Method=TIES2026.04 | 47.48 | |
| DARERank=32, Merge Method=Task Arithmetic2026.04 | 46.79 | |
| Core SpaceRank=8, Merge Method=Task Arithmetic2026.04 | 46.62 | |
| DELLARank=64, Merge Method=TIES2026.04 | 46.4 | |
| Core SpaceRank=64, Merge Method=Task Arithmetic2026.04 | 46.33 | |
| PicoRank=32, Merge Method=TSV-M2026.04 | 46.21 | |
| DELLARank=32, Merge Method=Task Arithmetic2026.04 | 45.05 | |
| DARERank=64, Merge Method=Task Arithmetic2026.04 | 45.04 | |
| DELLARank=8, Merge Method=Task Arithmetic2026.04 | 44.73 | |
| PicoRank=16, Merge Method=Task Arithmetic2026.04 | 44.3 | |
| KnOTSRank=8, Merge Method=Task Arithmetic2026.04 | 44.12 | |
| DARERank=8, Merge Method=Task Arithmetic2026.04 | 44.04 | |
| DELLARank=64, Merge Method=Task Arithmetic2026.04 | 44 | |
| Core SpaceRank=8, Merge Method=TIES2026.04 | 43.95 | |
| DELLARank=8, Merge Method=TIES2026.04 | 43.53 | |
| PicoRank=16, Merge Method=TIES2026.04 | 43.28 | |
| PicoRank=16, Merge Method=TSV-M2026.04 | 43.05 | |
| Task ArithmeticRank=8, Calibration=No Calibration2026.04 | 42.85 | |
| KnOTSRank=64, Merge Method=Task Arithmetic2026.04 | 42.7 | |
| Task ArithmeticRank=64, Calibration=No Calibration2026.04 | 42.57 | |
| KnOTSRank=16, Merge Method=TIES2026.04 | 42.07 | |
| DARERank=16, Merge Method=Task Arithmetic2026.04 | 41.86 | |
| Task ArithmeticRank=16, Calibration=No Calibration2026.04 | 40.93 | |
| DELLARank=16, Merge Method=Task Arithmetic2026.04 | 40.89 | |
| KnOTSRank=16, Merge Method=Task Arithmetic2026.04 | 40.88 | |
| Core SpaceRank=16, Merge Method=Task Arithmetic2026.04 | 40.65 | |
| PicoRank=64, Merge Method=TSV-M2026.04 | 39.76 | |
| TIESRank=16, Calibration=No Calibration2026.04 | 38.59 | |
| DARERank=16, Merge Method=TSV-M2026.04 | 37.22 | |
| DARERank=16, Merge Method=TIES2026.04 | 36.6 | |
| DELLARank=16, Merge Method=TSV-M2026.04 | 35.77 | |
| KnOTSRank=16, Merge Method=TSV-M2026.04 | 35.23 | |
| Core SpaceRank=16, Merge Method=TIES2026.04 | 35.17 | |
| TSV-MRank=16, Calibration=No Calibration2026.04 | 34.73 | |
| TSV-MRank=32, Calibration=No Calibration2026.04 | 32.06 | |
| KnOTSRank=32, Merge Method=TSV-M2026.04 | 31.73 | |
| TSV-MRank=8, Calibration=No Calibration2026.04 | 31.05 | |
| DELLARank=16, Merge Method=TIES2026.04 | 31.03 | |
| DELLARank=32, Merge Method=TSV-M2026.04 | 30.96 | |
| DARERank=64, Merge Method=TSV-M2026.04 | 30.52 | |
| DARERank=32, Merge Method=TSV-M2026.04 | 30.31 | |
| Core SpaceRank=64, Merge Method=TSV-M2026.04 | 30.11 | |
| Core SpaceRank=32, Merge Method=TSV-M2026.04 | 29.77 | |
| KnOTSRank=8, Merge Method=TSV-M2026.04 | 29.57 | |
| Core SpaceRank=16, Merge Method=TSV-M2026.04 | 29.45 | |
| TSV-MRank=64, Calibration=No Calibration2026.04 | 29.26 | |
| DELLARank=64, Merge Method=TSV-M2026.04 | 29.19 | |
| KnOTSRank=64, Merge Method=TSV-M2026.04 | 29.09 | |
| DARERank=8, Merge Method=TSV-M2026.04 | 28.11 | |
| Core SpaceRank=8, Merge Method=TSV-M2026.04 | 28.07 | |
| DELLARank=8, Merge Method=TSV-M2026.04 | 27.9 |