Mandate Tool Invocation on UPI-specific dataset Mandate flows
41.98Accuracy (Pause)Post-SFT
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Post-SFTTraining Stage=Supervised Fine-Tuning2026.02 | 41.98 | 76.24 | 77.42 | |
| Base ModelTraining Stage=Pre-SFT2026.02 | 3.27 | 40 | 46.15 |
Loading the SOTA2 catalog…