Jailbreak Defense on Behaviours (test)
0.9ASRVICUNA
Evaluation Results
| Method | Links | |
|---|---|---|
| VICUNASmooth q=5, Smoothing Type=Swap2024.10 | 0.9 | |
| PRO-VICUNA-HUBERSmooth q=1, delta=0.1, Smoothing Type=Swap2024.10 | 0.9 | |
| PRO-VICUNA-HUBERSmooth q=3, delta=0.1, Smoothing Type=Swap2024.10 | 0.9 | |
| PRO-VICUNA-HUBERSmooth q=5, delta=0.1, Smoothing Type=Swap2024.10 | 0.9 | |
| PRO-VICUNA-HUBERSmooth q=3, delta=0.2, Smoothing Type=Swap2024.10 | 0.9 | |
| PRO-VICUNA-HUBERSmooth q=5, delta=0.2, Smoothing Type=Swap2024.10 | 0.9 | |
| PRO-VICUNA-HUBERSmooth q=3, delta=0.3, Smoothing Type=Swap2024.10 | 0.9 | |
| PRO-VICUNA-HUBERSmooth q=5, delta=0.3, Smoothing Type=Swap2024.10 | 0.9 | |
| PRO-VICUNA-HUBERSmooth q=3, delta=0.4, Smoothing Type=Swap2024.10 | 0.9 | |
| PRO-VICUNA-HUBERSmooth q=5, delta=0.4, Smoothing Type=Swap2024.10 | 0.9 | |
| PRO-VICUNA-HUBERSmooth q=3, delta=0.5, Smoothing Type=Swap2024.10 | 0.9 | |
| PRO-VICUNA-HUBERSmooth q=5, delta=0.5, Smoothing Type=Swap2024.10 | 0.9 | |
| PRO-VICUNA-HUBERSmooth q=3, delta=0.8, Smoothing Type=Swap2024.10 | 0.9 | |
| PRO-VICUNA-HUBERSmooth q=5, delta=0.8, Smoothing Type=Swap2024.10 | 0.9 | |
| PRO-VICUNA-HUBERSmooth q=5, delta=1.0, Smoothing Type=Swap2024.10 | 0.9 | |
| PRO-VICUNA-HUBERSmooth q=0, delta=0.1, Smoothing Type=Swap2024.10 | 1.8 | |
| PRO-VICUNA-HUBERSmooth q=1, delta=0.2, Smoothing Type=Swap2024.10 | 1.8 | |
| PRO-VICUNA-HUBERSmooth q=5, delta=1.5, Smoothing Type=Swap2024.10 | 1.8 | |
| PRO-VICUNA-HUBERSmooth q=5, delta=2.0, Smoothing Type=Swap2024.10 | 1.8 | |
| PRO-VICUNA-HUBERSmooth q=1, delta=0.3, Smoothing Type=Swap2024.10 | 2.7 | |
| PRO-VICUNA-HUBERSmooth q=5, delta=3.0, Smoothing Type=Swap2024.10 | 7.3 | |
| PRO-VICUNA-HUBERSmooth q=0, delta=0.2, Smoothing Type=Swap2024.10 | 8.2 | |
| PRO-VICUNA-HUBERSmooth q=3, delta=1.0, Smoothing Type=Swap2024.10 | 11.8 | |
| PRO-VICUNA-HUBERSmooth q=3, delta=1.5, Smoothing Type=Swap2024.10 | 16.4 | |
| VICUNASmooth q=3, Smoothing Type=Swap2024.10 | 20.9 | |
| PRO-VICUNA-HUBERSmooth q=0, delta=0.3, Smoothing Type=Swap2024.10 | 21.8 | |
| PRO-VICUNA-HUBERSmooth q=1, delta=0.4, Smoothing Type=Swap2024.10 | 21.8 | |
| PRO-VICUNA-HUBERSmooth q=3, delta=2.0, Smoothing Type=Swap2024.10 | 21.8 | |
| PRO-VICUNA-HUBERSmooth q=1, delta=0.5, Smoothing Type=Swap2024.10 | 23.6 | |
| PRO-VICUNA-HUBERSmooth q=0, delta=0.4, Smoothing Type=Swap2024.10 | 30.9 | |
| PRO-VICUNA-HUBERSmooth q=3, delta=3.0, Smoothing Type=Swap2024.10 | 31.8 | |
| PRO-VICUNA-HUBERSmooth q=0, delta=0.5, Smoothing Type=Swap2024.10 | 36.4 | |
| PRO-VICUNA-HUBERSmooth q=1, delta=0.8, Smoothing Type=Swap2024.10 | 40 | |
| PRO-VICUNA-HUBERSmooth q=1, delta=1.0, Smoothing Type=Swap2024.10 | 53.6 | |
| PRO-VICUNA-HUBERSmooth q=1, delta=1.5, Smoothing Type=Swap2024.10 | 60 | |
| PRO-VICUNA-HUBERSmooth q=0, delta=0.8, Smoothing Type=Swap2024.10 | 61.8 | |
| PRO-VICUNA-HUBERSmooth q=0, delta=1.0, Smoothing Type=Swap2024.10 | 70 | |
| VICUNASmooth q=1, Smoothing Type=Swap2024.10 | 71.8 | |
| PRO-VICUNA-HUBERSmooth q=1, delta=2.0, Smoothing Type=Swap2024.10 | 73.6 | |
| PRO-VICUNA-HUBERSmooth q=0, delta=1.5, Smoothing Type=Swap2024.10 | 74.5 | |
| PRO-VICUNA-HUBERSmooth q=1, delta=3.0, Smoothing Type=Swap2024.10 | 74.5 | |
| PRO-VICUNA-HUBERSmooth q=0, delta=2.0, Smoothing Type=Swap2024.10 | 82.7 | |
| PRO-VICUNA-HUBERSmooth q=0, delta=3.0, Smoothing Type=Swap2024.10 | 90 | |
| VICUNASmooth q=0, Smoothing Type=Swap2024.10 | 91.8 |