List your company

Jailbreak Attack on Behaviours

0.9ASRPRO-VICUNA-HUBER
-2.73621.80746.3570.893Oct 30, 2024

Evaluation results

MethodLinks
PRO-VICUNA-HUBERSmooth q=1, delta=0.1, Smoothing Type=Patch2024.100.9
PRO-VICUNA-HUBERSmooth q=3, delta=0.1, Smoothing Type=Patch2024.100.9
PRO-VICUNA-HUBERSmooth q=5, delta=0.1, Smoothing Type=Patch2024.100.9
PRO-VICUNA-HUBERSmooth q=10, delta=0.1, Smoothing Type=Patch2024.100.9
PRO-VICUNA-HUBERSmooth q=15, delta=0.1, Smoothing Type=Patch2024.100.9
PRO-VICUNA-HUBERSmoothing Type=Insert, delta (δ)=0.1, Smooth q=12024.100.9
PRO-VICUNA-HUBERSmoothing Type=Insert, delta (δ)=0.1, Smooth q=32024.100.9
PRO-VICUNA-HUBERSmoothing Type=Insert, delta (δ)=0.1, Smooth q=52024.100.9
PRO-VICUNA-HUBERSmoothing Type=Insert, delta (δ)=0.1, Smooth q=102024.100.9
PRO-VICUNA-HUBERSmoothing Type=Insert, delta (δ)=0.1, Smooth q=152024.100.9
PRO-VICUNA-HUBERSmooth q=0, delta=0.1, Smoothing Type=Patch2024.101.8
VICUNASmoothing Type=Insert, Smooth q=152024.101.8
PRO-VICUNA-HUBERSmoothing Type=Insert, delta (δ)=0.1, Smooth q=02024.101.8
VICUNASmoothing Type=Insert, Smooth q=102024.104.5
VICUNASmoothing Type=Insert, Smooth q=52024.1010.9
VicunaSmooth q=15, Smoothing Type=Patch2024.1014.5
VicunaSmooth q=10, Smoothing Type=Patch2024.1021.8
VicunaSmooth q=5, Smoothing Type=Patch2024.1039.1
VICUNASmoothing Type=Insert, Smooth q=32024.1044.5
Pro-Vicuna (Best)Num of attack queries=122024.1050.7
Pro-Vicuna-Huber δ = 0.7Num of attack queries=122024.1050.7
Pro-Vicuna (Best)Num of attack queries=132024.1055.9
Pro-Vicuna-Huber δ = 0.7Num of attack queries=132024.1055.9
VicunaSmooth q=3, Smoothing Type=Patch2024.1057.3
Pro-Vicuna-Huber δ = 0.3Num of attack queries=122024.1060.2
Pro-Vicuna-Huber δ = 0.3Num of attack queries=132024.1060.2
Pro-Vicuna (Best)Num of attack queries=142024.1060.8
Pro-Vicuna-Huber δ = 0.5Num of attack queries=122024.1060.8
Pro-Vicuna-Huber δ = 0.7Num of attack queries=142024.1060.8
VicunaNum of attack queries=122024.1061.4
Pro-Vicuna-Huber δ = 0.5Num of attack queries=132024.1061.8
Pro-Vicuna-Huber δ = 0.5Num of attack queries=142024.1062.7
Pro-Vicuna (Best)Num of attack queries=152024.1064.3
Pro-Vicuna-Huber δ = 0.7Num of attack queries=152024.1064.3
Pro-Vicuna-Huber δ = 0.3Num of attack queries=142024.1065
VicunaNum of attack queries=132024.1065.2
Pro-Vicuna-Huber δ = 0.5Num of attack queries=152024.1066.7
Pro-Vicuna (Best)Num of attack queries=162024.1067.4
Pro-Vicuna-Huber δ = 0.7Num of attack queries=162024.1067.4
Pro-Vicuna-Huber δ = 0.5Num of attack queries=162024.1067.6
Pro-Vicuna (Best)Num of attack queries=172024.1070.5
Pro-Vicuna-Huber δ = 0.7Num of attack queries=172024.1070.5
Pro-Vicuna-Huber δ = 0.3Num of attack queries=152024.1070.9
VicunaNum of attack queries=142024.1071.5
Pro-Vicuna-Huber δ = 0.5Num of attack queries=172024.1071.6
VicunaSmooth q=1, Smoothing Type=Patch2024.1071.8
Pro-Vicuna-Huber δ = 0.3Num of attack queries=162024.1072.8
Pro-Vicuna (Best)Num of attack queries=182024.1074
Pro-Vicuna-Huber δ = 0.7Num of attack queries=182024.1074
Pro-Vicuna-Huber δ = 0.3Num of attack queries=172024.1075.7
VicunaNum of attack queries=152024.1075.8
Pro-Vicuna (Best)Num of attack queries=192024.1077.7
Pro-Vicuna-Huber δ = 0.3Num of attack queries=182024.1077.7
Pro-Vicuna-Huber δ = 0.3Num of attack queries=192024.1077.7
Pro-Vicuna-Huber δ = 0.5Num of attack queries=182024.1078.4
Pro-Vicuna (Best)Num of attack queries=202024.1078.6
Pro-Vicuna-Huber δ = 0.3Num of attack queries=202024.1078.6
VicunaNum of attack queries=162024.1078.7
VICUNASmoothing Type=Insert, Smooth q=12024.1079.1
Pro-Vicuna-Huber δ = 0.7Num of attack queries=192024.1079.7
Pro-Vicuna-Huber δ = 0.5Num of attack queries=192024.1081.4
Pro-Vicuna-Huber δ = 0.5Num of attack queries=202024.1082.4
Pro-Vicuna-Huber δ = 0.7Num of attack queries=202024.1082.4
VicunaNum of attack queries=172024.1082.6
VicunaNum of attack queries=182024.1084.1
VicunaNum of attack queries=192024.1086.5
VicunaNum of attack queries=202024.1087.4
VicunaSmooth q=0, Smoothing Type=Patch2024.1091.8
VICUNASmoothing Type=Insert, Smooth q=02024.1091.8