Binary Classification on FSFR
90.85AccuracyARSPO
Evaluation results
| Method | Links | |
|---|---|---|
| ARSPOTraining Split=FSFRrl, Base Model=SFT2026.02 | 90.85 | |
| SAPOTraining Split=FSFRrl, Base Model=SFT2026.02 | 90.75 | |
| GRPOTraining Split=FSFRrl, Base Model=SFT2026.02 | 89.42 | |
| Gemini3Model Category=SOTA MLLM2026.02 | 74.54 | |
| GPT5Model Category=SOTA MLLM2026.02 | 73.71 | |
| Seed1.6Model Category=SOTA MLLM2026.02 | 72.99 | |
| SFTTraining Split=FSFRsft2026.02 | 53.38 |