Mathematical Reasoning on MATHOAI (test)
76.2AccuracyGLORE
Evaluation results
| Method | Links | |
|---|---|---|
| GLOREBackbone=Qwen2.5-7B-Instruct2025.03 | 76.2 | |
| SFTBackbone=Qwen2.5-7B-Instruct2025.03 | 74.8 | |
| RoTBackbone=Qwen2.5-7B-Instruct2025.03 | 73.8 | |
| MathNeuroBackbone=Qwen2.5-7B-Instruct2025.03 | 73.6 | |
| Zero-shot CoTBackbone=Qwen2.5-7B-Instruct2025.03 | 72.8 | |
| BoostStepBackbone=Qwen2.5-7B-Instruct2025.03 | 70.8 | |
| Few-shot CoTBackbone=Qwen2.5-7B-Instruct2025.03 | 69.8 | |
| GLOREBackbone=Llama3.1-8B-Instruct2025.03 | 51.6 | |
| SFTBackbone=Llama3.1-8B-Instruct2025.03 | 50.2 | |
| RoTBackbone=Llama3.1-8B-Instruct2025.03 | 49.4 | |
| MathNeuroBackbone=Llama3.1-8B-Instruct2025.03 | 49 | |
| Zero-shot CoTBackbone=Llama3.1-8B-Instruct2025.03 | 48.2 | |
| BoostStepBackbone=Llama3.1-8B-Instruct2025.03 | 47.6 | |
| Few-shot CoTBackbone=Llama3.1-8B-Instruct2025.03 | 45.6 |