Automated Short Answer Scoring on SAS Geography subject (OOD)
43.3AccuracyEDIT-SFT
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| EDIT-SFTCategory=EDIT (Ours)2026.06 | 43.3 | 63.3 | 1.267 | 0.779 | |
| EDIT-RLCategory=EDIT (Ours)2026.06 | 43.3 | 60 | 1.3 | 0.756 | |
| BaseCategory=Base2026.06 | 41.4 | 55.2 | 1.483 | 0.706 | |
| GRPOCategory=RL-only2026.06 | 40 | 53.3 | 1.533 | 0.676 | |
| DGPOCategory=RL-only2026.06 | 36.7 | 60 | 1.533 | 0.685 | |
| SFTCategory=InT2026.06 | 36.7 | 60 | 1.533 | 0.745 | |
| RLCategory=InT2026.06 | 33.3 | 53.3 | 1.667 | 0.675 |