Binary Classification on AGIN-Rat (test)
82.71Macro-F1RL-Single
Evaluation Results
| Method | Links | |
|---|---|---|
| RL-Singletraining=Single-task RL2026.03 | 82.71 | |
| MT-RL-Judgetraining=Multi-task RL2026.03 | 81.58 | |
| SFT-Unifiedtraining=Unified multi-task SFT2026.03 | 81.31 | |
| SFT-Singletraining=Single-task SFT2026.03 | 78.08 | |
| Off-the-shelf2026.03 | 64.77 |