Long-form QA on WaterBench (test)
24.06GM ScoreSWEET
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| SWEETModel=OPT-1.3B2025.12 | 24.06 | 1 | 1 | |
| UnbiasedModel=OPT-1.3B2025.12 | 23.99 | 0.96 | 0.985 | |
| SynthIDModel=OPT-1.3B2025.12 | 23.93 | 0.95 | 0.965 | |
| DIPmarkModel=OPT-1.3B2025.12 | 23.92 | 0.885 | 0.91 | |
| OriginalModel=OPT-1.3B2025.12 | 23.87 | — | — | |
| KGWModel=OPT-1.3B2025.12 | 23.86 | 0.99 | 0.98 | |
| EWDModel=OPT-1.3B2025.12 | 23.69 | 1 | 1 | |
| DualGuardModel=OPT-1.3B2025.12 | 23.35 | 0.985 | 0.985 | |
| AARModel=OPT-1.3B2025.12 | 23.19 | 0.985 | 0.98 | |
| XSIRModel=OPT-1.3B2025.12 | 22.52 | 0.9 | 0.84 | |
| SIRModel=OPT-1.3B2025.12 | 22.34 | 1 | 0.925 |