List your company

Multi-domain Evaluation (Math, Code, Finance, Medical) Aggregate

30.65Math AccuracyPico
18.825221.895124.96528.0349Apr 18, 2026

Evaluation results

MethodLinks
PicoMerger=Task Arithmetic2026.0430.6533.7254.3158.5444.3
Pico w/ Task ArithmeticCalibration=Pico, Merging Rule=Task Arithmetic2026.0430.6533.7254.3158.5444.3
PicoMerger=TIES2026.0430.6130.3650.9661.1643.28
DAREMerger=TSV-M2026.0429.2616.244.8158.5937.22
PicoMerger=TSV-M2026.0428.8329.4252.5761.3943.05
KnOTSMerger=TIES2026.0428.7630.1252.7756.6142.07
Math LoRATraining Type=Single-Domain Adapter2026.0428.310.944.0336.6529.97
No CalibrationMerger=TIES2026.0428.0821.1752.7152.4138.59
DELLAMerger=TSV-M2026.0427.7815.9846.6252.6835.77
DAREMerger=Task Arithmetic2026.0427.5729.3152.0658.5141.86
KnOTSMerger=Task Arithmetic2026.0427.2727.1451.2757.8440.88
No CalibrationMerger=Task Arithmetic2026.0426.6628.2350.4458.3940.93
KnOTSMerger=TSV-M2026.0426.6615.8246.2552.1835.23
DELLAMerger=Task Arithmetic2026.0426.5126.6251.8758.5740.89
No CalibrationMerger=TSV-M2026.0426.3615.744.8751.9734.73
Core SpaceMerger=Task Arithmetic2026.0426.227.7349.8758.840.65
DAREMerger=TIES2026.0426.0119.3651.349.7436.6
Joint Multi-Task LoRATraining Type=Jointly Trained Multi-Task2026.0425.4431.5146.5544.0536.88
Coding LoRATraining Type=Single-Domain Adapter2026.0425.0535.9848.1255.1641.08
Finance LoRATraining Type=Single-Domain Adapter2026.0424.7714.454.8449.7635.94
Medical LoRATraining Type=Single-Domain Adapter2026.0424.3831.948.4859.541.06
Core SpaceMerger=TIES2026.0424.1324.942.349.3635.17
Core SpaceMerger=TSV-M2026.0423.8514.1535.6844.1329.45
DELLAMerger=TIES2026.0419.2817.0947.540.2531.03