Toxicity Detection on Toxicity Detection 64 model-persona combinations (8 models x 8 personas)
58Win CountSVM Meta-Ensemble
Evaluation Results
| Method | Links | |
|---|---|---|
| SVM Meta-EnsembleComparison Baseline=Default2026.01 | 58 | |
| Best MajorityComparison Baseline=Default2026.01 | 57 | |
| SVM Meta-EnsembleComparison Baseline=Persona Prompt2026.01 | 51 | |
| Weighted MajorityComparison Baseline=Default2026.01 | 50 | |
| Theoretical Optimal Weighted MajorityComparison Baseline=Default2026.01 | 50 | |
| SVM Meta-EnsembleComparison Baseline=Value Profile2026.01 | 47 | |
| SVM Meta-EnsembleComparison Baseline=Persona Optimized Prompt2026.01 | 44 | |
| Persona Optimized PromptComparison Baseline=Default2026.01 | 43 | |
| Value ProfileComparison Baseline=Default2026.01 | 42 | |
| SVM Meta-EnsembleComparison Baseline=Theoretical Optimal Weighted Majority2026.01 | 41 | |
| Best MajorityComparison Baseline=Persona Prompt2026.01 | 40 | |
| SVM Meta-EnsembleComparison Baseline=Weighted Majority2026.01 | 40 | |
| Weighted MajorityComparison Baseline=Persona Prompt2026.01 | 37 | |
| Theoretical Optimal Weighted MajorityComparison Baseline=Persona Prompt2026.01 | 36 | |
| Persona PromptComparison Baseline=Default2026.01 | 34 | |
| SVM Meta-EnsembleComparison Baseline=Best Majority2026.01 | 29 | |
| Best MajorityComparison Baseline=Persona Optimized Prompt2026.01 | 28 | |
| Value ProfileComparison Baseline=Persona Prompt2026.01 | 25 | |
| Weighted MajorityComparison Baseline=Value Profile2026.01 | 25 | |
| Best MajorityComparison Baseline=Value Profile2026.01 | 24 | |
| Persona Optimized PromptComparison Baseline=Persona Prompt2026.01 | 23 | |
| Theoretical Optimal Weighted MajorityComparison Baseline=Value Profile2026.01 | 23 | |
| Weighted MajorityComparison Baseline=Persona Optimized Prompt2026.01 | 23 | |
| Theoretical Optimal Weighted MajorityComparison Baseline=Persona Optimized Prompt2026.01 | 23 | |
| Best MajorityComparison Baseline=Theoretical Optimal Weighted Majority2026.01 | 21 | |
| Best MajorityComparison Baseline=Weighted Majority2026.01 | 20 | |
| Persona Optimized PromptComparison Baseline=Value Profile2026.01 | 19 | |
| Value ProfileComparison Baseline=Persona Optimized Prompt2026.01 | 18 | |
| Persona PromptComparison Baseline=Value Profile2026.01 | 13 | |
| Persona PromptComparison Baseline=Persona Optimized Prompt2026.01 | 12 | |
| Value ProfileComparison Baseline=Weighted Majority2026.01 | 10 | |
| Value ProfileComparison Baseline=Theoretical Optimal Weighted Majority2026.01 | 10 | |
| DefaultComparison Baseline=Value Profile2026.01 | 8 | |
| DefaultComparison Baseline=Persona Prompt2026.01 | 7 | |
| Persona Optimized PromptComparison Baseline=Theoretical Optimal Weighted Majority2026.01 | 6 | |
| DefaultComparison Baseline=Persona Optimized Prompt2026.01 | 5 | |
| Persona PromptComparison Baseline=Weighted Majority2026.01 | 4 | |
| Persona PromptComparison Baseline=Theoretical Optimal Weighted Majority2026.01 | 4 | |
| Persona Optimized PromptComparison Baseline=Weighted Majority2026.01 | 4 | |
| Weighted MajorityComparison Baseline=Theoretical Optimal Weighted Majority2026.01 | 3 | |
| Weighted MajorityComparison Baseline=SVM Meta-Ensemble2026.01 | 3 | |
| Theoretical Optimal Weighted MajorityComparison Baseline=SVM Meta-Ensemble2026.01 | 3 | |
| Weighted MajorityComparison Baseline=Best Majority2026.01 | 2 | |
| Best MajorityComparison Baseline=SVM Meta-Ensemble2026.01 | 2 | |
| Persona PromptComparison Baseline=SVM Meta-Ensemble2026.01 | 1 | |
| Persona Optimized PromptComparison Baseline=SVM Meta-Ensemble2026.01 | 1 | |
| Theoretical Optimal Weighted MajorityComparison Baseline=Best Majority2026.01 | 1 | |
| DefaultComparison Baseline=Weighted Majority2026.01 | 0 | |
| DefaultComparison Baseline=Best Majority2026.01 | 0 | |
| DefaultComparison Baseline=Theoretical Optimal Weighted Majority2026.01 | 0 | |
| DefaultComparison Baseline=SVM Meta-Ensemble2026.01 | 0 | |
| Persona PromptComparison Baseline=Best Majority2026.01 | 0 | |
| Value ProfileComparison Baseline=Best Majority2026.01 | 0 | |
| Value ProfileComparison Baseline=SVM Meta-Ensemble2026.01 | 0 | |
| Persona Optimized PromptComparison Baseline=Best Majority2026.01 | 0 | |
| Theoretical Optimal Weighted MajorityComparison Baseline=Weighted Majority2026.01 | 0 |