Uncertainty Quantification on TQA 4 additional samples (test)
0.937AUROCClustered Self-Assessment
Evaluation Results
| Method | Links | |
|---|---|---|
| Clustered Self-AssessmentModel=Qwen2.5-32B2026.06 | 0.937 | |
| Clustered Self-AssessmentModel=Qwen2.5-14B2026.06 | 0.936 | |
| Clustered Self-AssessmentModel=Qwen3-14B2026.06 | 0.932 | |
| Clustered Self-AssessmentModel=Qwen3-30B-A3B2026.06 | 0.929 | |
| Clustered Self-AssessmentModel=Gemma-3-12B2026.06 | 0.927 | |
| Clustered Self-AssessmentModel=Gemma-3-27B2026.06 | 0.922 | |
| Clustered Self-AssessmentModel=Qwen2.5-7B2026.06 | 0.921 | |
| SARModel=Gemma-3-12B2026.06 | 0.904 | |
| DegModel=Gemma-3-12B2026.06 | 0.898 | |
| ProbabilityModel=Qwen3-14B2026.06 | 0.897 | |
| tokenSARModel=Gemma-3-12B2026.06 | 0.897 | |
| sentenceSARModel=Qwen3-14B2026.06 | 0.897 | |
| PerplexityModel=Qwen3-14B2026.06 | 0.895 | |
| PerplexityModel=Gemma-3-12B2026.06 | 0.895 | |
| LN EntropyModel=Qwen3-14B2026.06 | 0.895 | |
| PerplexityModel=Gemma-3-27B2026.06 | 0.894 | |
| Predictive EntropyModel=Qwen3-14B2026.06 | 0.894 | |
| SARModel=Gemma-3-27B2026.06 | 0.892 | |
| tokenSARModel=Gemma-3-27B2026.06 | 0.891 | |
| ProbabilityModel=Qwen2.5-7B2026.06 | 0.89 | |
| P(True)Model=Qwen2.5-32B2026.06 | 0.89 | |
| SARModel=Qwen3-14B2026.06 | 0.89 | |
| EigVModel=Gemma-3-12B2026.06 | 0.889 | |
| PerplexityModel=Qwen2.5-7B2026.06 | 0.888 | |
| sentenceSARModel=Qwen2.5-7B2026.06 | 0.887 | |
| LN EntropyModel=Qwen2.5-7B2026.06 | 0.886 | |
| P(True)Model=Qwen3-14B2026.06 | 0.886 | |
| Predictive EntropyModel=Qwen2.5-7B2026.06 | 0.885 | |
| LN EntropyModel=Gemma-3-12B2026.06 | 0.885 | |
| sentenceSARModel=Qwen2.5-32B2026.06 | 0.884 | |
| ProbabilityModel=Qwen2.5-32B2026.06 | 0.883 | |
| ProbabilityModel=Gemma-3-27B2026.06 | 0.883 | |
| ProbabilityModel=Qwen2.5-14B2026.06 | 0.882 | |
| LN EntropyModel=Qwen2.5-32B2026.06 | 0.881 | |
| sentenceSARModel=Qwen2.5-14B2026.06 | 0.881 | |
| DegModel=Gemma-3-27B2026.06 | 0.881 | |
| PerplexityModel=Qwen2.5-14B2026.06 | 0.88 | |
| Predictive EntropyModel=Qwen2.5-32B2026.06 | 0.88 | |
| SARModel=Qwen2.5-7B2026.06 | 0.88 | |
| PerplexityModel=Qwen2.5-32B2026.06 | 0.879 | |
| LN EntropyModel=Qwen2.5-14B2026.06 | 0.879 | |
| sentenceSARModel=Gemma-3-12B2026.06 | 0.879 | |
| ProbabilityModel=Gemma-3-12B2026.06 | 0.878 | |
| Predictive EntropyModel=Qwen2.5-14B2026.06 | 0.878 | |
| EigVModel=Gemma-3-27B2026.06 | 0.878 | |
| SARModel=Qwen2.5-32B2026.06 | 0.877 | |
| PerplexityModel=Qwen3-30B-A3B2026.06 | 0.876 | |
| P(True)Model=Qwen3-30B-A3B2026.06 | 0.876 | |
| sentenceSARModel=Gemma-3-27B2026.06 | 0.876 | |
| ProbabilityModel=Qwen3-30B-A3B2026.06 | 0.875 | |
| SARModel=Qwen2.5-14B2026.06 | 0.875 | |
| LN EntropyModel=Qwen3-30B-A3B2026.06 | 0.874 | |
| SARModel=Qwen3-30B-A3B2026.06 | 0.874 | |
| sentenceSARModel=Qwen3-30B-A3B2026.06 | 0.873 | |
| LN EntropyModel=Gemma-3-27B2026.06 | 0.87 | |
| Predictive EntropyModel=Qwen3-30B-A3B2026.06 | 0.869 | |
| Predictive EntropyModel=Gemma-3-12B2026.06 | 0.869 | |
| P(True)Model=Qwen2.5-14B2026.06 | 0.868 | |
| tokenSARModel=Qwen3-14B2026.06 | 0.868 | |
| Predictive EntropyModel=Gemma-3-27B2026.06 | 0.865 | |
| Semantic EntropyModel=Gemma-3-12B2026.06 | 0.859 | |
| tokenSARModel=Qwen2.5-7B2026.06 | 0.857 | |
| Semantic EntropyModel=Gemma-3-27B2026.06 | 0.855 | |
| tokenSARModel=Qwen2.5-14B2026.06 | 0.853 | |
| EigVModel=Qwen3-14B2026.06 | 0.852 | |
| tokenSARModel=Qwen3-30B-A3B2026.06 | 0.851 | |
| tokenSARModel=Qwen2.5-32B2026.06 | 0.846 | |
| EigVModel=Qwen2.5-7B2026.06 | 0.841 | |
| EigVModel=Qwen2.5-32B2026.06 | 0.841 | |
| Semantic EntropyModel=Qwen3-14B2026.06 | 0.839 | |
| DegModel=Qwen3-14B2026.06 | 0.839 | |
| EccModel=Gemma-3-27B2026.06 | 0.837 | |
| EccModel=Gemma-3-12B2026.06 | 0.834 | |
| P(True)Model=Qwen2.5-7B2026.06 | 0.832 | |
| Semantic EntropyModel=Qwen2.5-7B2026.06 | 0.829 | |
| DegModel=Qwen2.5-7B2026.06 | 0.828 | |
| DegModel=Qwen2.5-32B2026.06 | 0.828 | |
| EigVModel=Qwen3-30B-A3B2026.06 | 0.827 | |
| Semantic EntropyModel=Qwen2.5-32B2026.06 | 0.826 | |
| Eigen ScoreModel=Qwen3-14B2026.06 | 0.82 | |
| EigVModel=Qwen2.5-14B2026.06 | 0.82 | |
| Semantic EntropyModel=Qwen3-30B-A3B2026.06 | 0.817 | |
| Eigen ScoreModel=Qwen2.5-7B2026.06 | 0.816 | |
| Eigen ScoreModel=Gemma-3-12B2026.06 | 0.814 | |
| EccModel=Qwen3-14B2026.06 | 0.814 | |
| DegModel=Qwen3-30B-A3B2026.06 | 0.813 | |
| Eigen ScoreModel=Qwen2.5-32B2026.06 | 0.81 | |
| DegModel=Qwen2.5-14B2026.06 | 0.809 | |
| Semantic EntropyModel=Qwen2.5-14B2026.06 | 0.807 | |
| Eigen ScoreModel=Gemma-3-27B2026.06 | 0.805 | |
| EccModel=Qwen2.5-7B2026.06 | 0.805 | |
| EccModel=Qwen2.5-32B2026.06 | 0.801 | |
| Eigen ScoreModel=Qwen3-30B-A3B2026.06 | 0.798 | |
| Eigen ScoreModel=Qwen2.5-14B2026.06 | 0.796 | |
| P(True)Model=Gemma-3-27B2026.06 | 0.789 | |
| EccModel=Qwen2.5-14B2026.06 | 0.785 | |
| EccModel=Qwen3-30B-A3B2026.06 | 0.785 | |
| P(True)Model=Gemma-3-12B2026.06 | 0.737 |