Uncertainty Calibration on MATH 500
0.861AUROCET-PE
Evaluation Results
| Method | Links | |
|---|---|---|
| ET-PEBackbone=Qwen2.5-14B-Instruct2026.01 | 0.861 | |
| PEBackbone=Qwen2.5-14B-Instruct2026.01 | 0.859 | |
| LSBackbone=Qwen2.5-14B-Instruct2026.01 | 0.852 | |
| ET-PEBackbone=Qwen2.5-7B-Instruct2026.01 | 0.84 | |
| PEBackbone=Qwen2.5-7B-Instruct2026.01 | 0.806 | |
| ET-PEBackbone=Qwen2.5-32B-Instruct2026.01 | 0.751 | |
| PEBackbone=Qwen2.5-32B-Instruct2026.01 | 0.751 | |
| LSBackbone=Qwen2.5-32B-Instruct2026.01 | 0.751 | |
| p(True)Backbone=Qwen2.5-7B-Instruct2026.01 | 0.736 | |
| LSBackbone=Qwen2.5-7B-Instruct2026.01 | 0.704 | |
| p(True)Backbone=Qwen2.5-32B-Instruct2026.01 | 0.694 | |
| LN-PEBackbone=Qwen2.5-14B-Instruct2026.01 | 0.693 | |
| SEBackbone=Qwen2.5-14B-Instruct2026.01 | 0.661 | |
| p(True)Backbone=Qwen2.5-14B-Instruct2026.01 | 0.657 | |
| LN-PEBackbone=Qwen2.5-7B-Instruct2026.01 | 0.606 | |
| LN-PEBackbone=Qwen2.5-32B-Instruct2026.01 | 0.599 | |
| SEBackbone=Qwen2.5-32B-Instruct2026.01 | 0.578 | |
| SEBackbone=Qwen2.5-7B-Instruct2026.01 | 0.558 |