Mathematical Reasoning on MathVerse-Plus All 1.0 (test)
43.6AccuracySpatialMath
Evaluation Results
| Method | Links | |
|---|---|---|
| SpatialMathBase Model=Phi-42026.01 | 43.6 | |
| SFTBase Model=Phi-42026.01 | 33.4 | |
| ICL (1-shot)Base Model=Phi-42026.01 | 28.4 | |
| Zero-ShotBase Model=Phi-42026.01 | 28.2 | |
| CoT (1-shot)Base Model=Phi-42026.01 | 28 | |
| SFT + Data AugmentationBase Model=Phi-42026.01 | 24.6 | |
| SpatialMathBase Model=LLaVA-NeXT-34B2026.01 | 23 | |
| SFT + Data AugmentationBase Model=LLaVA-NeXT-34B2026.01 | 19.2 | |
| Zero-ShotBase Model=LLaVA-NeXT-34B2026.01 | 18 | |
| SFTBase Model=LLaVA-NeXT-34B2026.01 | 18 | |
| CoT (1-shot)Base Model=LLaVA-NeXT-34B2026.01 | 17.2 | |
| ICL (1-shot)Base Model=LLaVA-NeXT-34B2026.01 | 10 |