Uncertainty Calibration on TQA
0.0721Brier ScoreClustered Self-Assessment
Evaluation Results
| Method | Links | |
|---|---|---|
| Clustered Self-AssessmentBackbone=Gemma-3-27B2026.06 | 0.0721 | |
| Clustered Self-AssessmentBackbone=Qwen2.5-32B2026.06 | 0.0843 | |
| Clustered Self-AssessmentBackbone=Qwen3-30B-A3B2026.06 | 0.0894 | |
| Clustered Self-AssessmentBackbone=Qwen2.5-14B2026.06 | 0.0917 | |
| Clustered Self-AssessmentBackbone=Qwen3-14B2026.06 | 0.0921 | |
| NSEBackbone=Gemma-3-27B2026.06 | 0.0937 | |
| Clustered Self-AssessmentBackbone=Gemma-3-12B2026.06 | 0.0943 | |
| P(True)Backbone=Qwen2.5-32B2026.06 | 0.1172 | |
| Clustered Self-AssessmentBackbone=Qwen2.5-7B2026.06 | 0.1183 | |
| NSEBackbone=Qwen2.5-32B2026.06 | 0.12 | |
| NSEBackbone=Qwen3-14B2026.06 | 0.1283 | |
| NSEBackbone=Qwen3-30B-A3B2026.06 | 0.1289 | |
| P(True)Backbone=Qwen3-14B2026.06 | 0.1315 | |
| NSEBackbone=Qwen2.5-14B2026.06 | 0.1342 | |
| P(True)Backbone=Qwen3-30B-A3B2026.06 | 0.1398 | |
| P(True)Backbone=Qwen2.5-14B2026.06 | 0.14 | |
| NSEBackbone=Gemma-3-12B2026.06 | 0.1421 | |
| NSEBackbone=Qwen2.5-7B2026.06 | 0.158 | |
| ProbabilityBackbone=Qwen2.5-7B2026.06 | 0.1674 | |
| P(True)Backbone=Qwen2.5-7B2026.06 | 0.1708 | |
| P(True)Backbone=Gemma-3-27B2026.06 | 0.1758 | |
| ProbabilityBackbone=Qwen3-14B2026.06 | 0.1807 | |
| ProbabilityBackbone=Qwen3-30B-A3B2026.06 | 0.1917 | |
| ProbabilityBackbone=Gemma-3-27B2026.06 | 0.1975 | |
| ProbabilityBackbone=Qwen2.5-14B2026.06 | 0.2001 | |
| P(True)Backbone=Gemma-3-12B2026.06 | 0.2147 | |
| ProbabilityBackbone=Qwen2.5-32B2026.06 | 0.2267 | |
| ProbabilityBackbone=Gemma-3-12B2026.06 | 0.2402 |