Mathematical Reasoning on Numina
45.5AccuracyAcquisitionSynthesis (Confidence)
Evaluation Results
| Method | Links | |
|---|---|---|
| AcquisitionSynthesis (Confidence)Model=Llama 8B, Category=AS2026.05 | 45.5 | |
| FilteredModel=Llama 8B, Category=Selection2026.05 | 45.2 | |
| AcquisitionSynthesis (Diversity)Model=Qwen 7B, Category=AS2026.05 | 45.2 | |
| Prismatic Syn.Model=Qwen 3B, Category=Synthesis2026.05 | 44.6 | |
| OriginalModel=Llama 8B, Category=Synthesis2026.05 | 44.5 | |
| RandomModel=Llama 8B, Category=Selection2026.05 | 43.7 | |
| AcquisitionSynthesis (Gradient Norm)Model=Llama 8B, Category=AS2026.05 | 43.7 | |
| AcquisitionSynthesis (Answer Variance)Model=Qwen 7B, Category=AS2026.05 | 43.4 | |
| AcquisitionSynthesis (Answer Variance)Model=Llama 8B, Category=AS2026.05 | 43.2 | |
| AcquisitionSynthesis (Format)Model=Qwen 7B, Category=AS2026.05 | 43.2 | |
| AcquisitionSynthesis (Format)Model=Llama 8B, Category=AS2026.05 | 43.1 | |
| AcquisitionSynthesis (Confidence)Model=Qwen 7B, Category=AS2026.05 | 43.1 | |
| AcquisitionSynthesis (Diversity)Model=Llama 8B, Category=AS2026.05 | 42.8 | |
| AcquisitionSynthesis (Gradient Norm)Model=Qwen 7B, Category=AS2026.05 | 42.5 | |
| AcquisitionSynthesis (Proximity)Model=Llama 8B, Category=AS2026.05 | 40.7 | |
| AcquisitionSynthesis (Proximity)Model=Qwen 7B, Category=AS2026.05 | 40.2 | |
| OriginalModel=Qwen 7B, Category=Synthesis2026.05 | 36.5 | |
| FilteredModel=Qwen 7B, Category=Selection2026.05 | 34.8 | |
| Prismatic Syn.Model=Qwen 7B, Category=Synthesis2026.05 | 34.8 | |
| AcquisitionSynthesis (Confidence)Model=Qwen 3B, Category=AS2026.05 | 33.7 | |
| Prismatic Syn.Model=Llama 8B, Category=Synthesis2026.05 | 33 | |
| DataEnvGymModel=Qwen 3B, Category=Synthesis2026.05 | 32.3 | |
| Untrained modelModel=Llama 8B2026.05 | 32.2 | |
| AcquisitionSynthesis (Answer Variance)Model=Qwen 3B, Category=AS2026.05 | 30.5 | |
| AcquisitionSynthesis (Gradient Norm)Model=Qwen 3B, Category=AS2026.05 | 29.7 | |
| AcquisitionSynthesis (Format)Model=Qwen 3B, Category=AS2026.05 | 29.6 | |
| OriginalModel=Qwen 3B, Category=Synthesis2026.05 | 29.4 | |
| Untrained modelModel=Qwen 7B2026.05 | 29.3 | |
| FilteredModel=Qwen 3B, Category=Selection2026.05 | 29.2 | |
| RandomModel=Qwen 3B, Category=Selection2026.05 | 28.8 | |
| Untrained modelModel=Qwen 3B2026.05 | 28.3 | |
| DataEnvGymModel=Llama 8B, Category=Synthesis2026.05 | 26.8 | |
| AcquisitionSynthesis (Diversity)Model=Qwen 3B, Category=AS2026.05 | 24.7 | |
| RandomModel=Qwen 7B, Category=Selection2026.05 | 22.9 | |
| AcquisitionSynthesis (Proximity)Model=Qwen 3B, Category=AS2026.05 | 22.1 | |
| DataEnvGymModel=Qwen 7B, Category=Synthesis2026.05 | 17.8 |