Binary Classification on AGIN-Tech (test)
81.37Macro F1 ScoreMT-RL-Judge
Evaluation Results
| Method | Links | |
|---|---|---|
| MT-RL-Judgetraining=Multi-task RL2026.03 | 81.37 | |
| SFT-Unifiedtraining=Unified multi-task SFT2026.03 | 81.22 | |
| RL-Singletraining=Single-task RL2026.03 | 80.77 | |
| SFT-Singletraining=Single-task SFT2026.03 | 77.04 | |
| Off-the-shelf2026.03 | 63.24 |