Conformal Hallucination Detection on Diverse Benchmarks Aggregated (test)
98.5R-CovACSE
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| ACSEalpha=0.05 (Strict)2026.05 | 98.5 | 1.32 | 2.4 | 97.1 | 55.4 | 4.4 | |
| CAPalpha=0.05 (Strict)2026.05 | 95.3 | 1.72 | 3.8 | 95.1 | 46.1 | 4.6 | |
| DDCRP-CPalpha=0.05 (Strict)2026.05 | 94.1 | 1.95 | 4.6 | 94.2 | 42.5 | 3.8 | |
| ACSEalpha=0.10 (Standard)2026.05 | 93.9 | 1.07 | 3 | 92.3 | 75.8 | 8.8 | |
| DDCRP-CPalpha=0.10 (Standard)2026.05 | 92.3 | 1.45 | 5.1 | 88.4 | 62.8 | 7.6 | |
| CAPalpha=0.10 (Standard)2026.05 | 90.2 | 1.32 | 4.2 | 91.7 | 65.4 | 9.3 | |
| ACSEalpha=0.20 (Permissive)2026.05 | 84.2 | 1.08 | 4.5 | 83.5 | 89.4 | 17.8 | |
| CAPalpha=0.20 (Permissive)2026.05 | 82.7 | 1.15 | 6.2 | 81.4 | 82.5 | 18.8 | |
| SUalpha=0.05 (Strict)2026.05 | 82.3 | 1.05 | 8.1 | 81 | 34.2 | 6.4 | |
| SUalpha=0.10 (Standard)2026.05 | 81.6 | 1.02 | 8.9 | 82 | 52.1 | 11.8 | |
| DDCRP-CPalpha=0.20 (Permissive)2026.05 | 81 | 1.22 | 7.4 | 79.1 | 79.1 | 15.2 | |
| SUalpha=0.20 (Permissive)2026.05 | 80.8 | 1.01 | 11.5 | 85.2 | 68.4 | 22.5 |