Safety Defense on Alpaca harmful subset (test)
26.6Harmful ScoreVaccine-SFT
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Vaccine-SFTharmful_ratio (p)=clean2024.05 | 26.6 | 95.3 | |
| SFTharmful_ratio (p)=clean2024.05 | 34.6 | 95.3 | |
| Lisaharmful_ratio (p)=clean2024.05 | 34.9 | 95.07 | |
| Lisaharmful_ratio (p)=0.12024.05 | 40.2 | 94.84 | |
| Lisaharmful_ratio (p)=0.32024.05 | 43.6 | 94.04 | |
| Lisaharmful_ratio (p)=12024.05 | 44.1 | 13.65 | |
| Lisaharmful_ratio (p)=0.52024.05 | 44.4 | 88.65 | |
| Lisaharmful_ratio (p)=0.72024.05 | 45.1 | 89.11 | |
| Lisaharmful_ratio (p)=0.92024.05 | 45.3 | 88.07 | |
| Vaccine-SFTharmful_ratio (p)=0.52024.05 | 51.4 | 81.88 | |
| SFTharmful_ratio (p)=0.12024.05 | 51.6 | 94.95 | |
| SFTharmful_ratio (p)=0.72024.05 | 51.7 | 89.33 | |
| SFTharmful_ratio (p)=0.52024.05 | 52.2 | 90.94 | |
| Vaccine-SFTharmful_ratio (p)=0.72024.05 | 52.5 | 67.78 | |
| SFTharmful_ratio (p)=12024.05 | 52.6 | 20.87 | |
| Vaccine-SFTharmful_ratio (p)=0.92024.05 | 52.6 | 82.11 | |
| Vaccine-SFTharmful_ratio (p)=0.12024.05 | 52.7 | 94.27 | |
| Vaccine-SFTharmful_ratio (p)=0.32024.05 | 53 | 94.38 | |
| SFTharmful_ratio (p)=0.92024.05 | 53.2 | 89.68 | |
| Vaccine-SFTharmful_ratio (p)=12024.05 | 53.4 | 0.11 | |
| SFTharmful_ratio (p)=0.32024.05 | 53.6 | 95.3 |