Mathematical Reasoning on SVAMP (full set)
71.1Exact Match AccuracyGCE
Evaluation Results
| Method | Links | |
|---|---|---|
| GCEModel=Qwen2.5, Params=1.5B2026.06 | 71.1 | |
| SMMDModel=Qwen2.5, Params=1.5B2026.06 | 71.1 | |
| CEModel=Qwen2.5, Params=1.5B2026.06 | 68.6 | |
| NTILModel=Qwen2.5, Params=1.5B2026.06 | 67.1 | |
| SMMDModel=SmolLM3, Params=3B2026.06 | 67.1 | |
| CEModel=SmolLM3, Params=3B2026.06 | 65.8 | |
| NTILModel=SmolLM3, Params=3B2026.06 | 65.8 | |
| SMMDModel=Llama3, Params=8B2026.06 | 65.3 | |
| GCEModel=SmolLM3, Params=3B2026.06 | 63.7 | |
| CEModel=Llama3, Params=8B2026.06 | 62.2 | |
| NTLModel=SmolLM3, Params=3B2026.06 | 60.5 | |
| NTLModel=Qwen2.5, Params=1.5B2026.06 | 59.8 | |
| NTILModel=Llama3, Params=8B2026.06 | 59.5 | |
| GCEModel=Llama3, Params=8B2026.06 | 58.4 | |
| NTLModel=Llama3, Params=8B2026.06 | 57.5 | |
| SMMDModel=Qwen2.5, Params=0.5B2026.06 | 43.9 | |
| NTLModel=Qwen2.5, Params=0.5B2026.06 | 43.2 | |
| NTILModel=Qwen2.5, Params=0.5B2026.06 | 40.6 | |
| CEModel=Qwen2.5, Params=0.5B2026.06 | 39.8 | |
| GCEModel=Qwen2.5, Params=0.5B2026.06 | 37.9 |