List your company
ResearchFormerly WizWand

Defense against adversarial prompts on Sneaky (test)

45.16Attack Success RateSDv1.5
-1.806410.386822.5834.7732Jul 1, 2026
Updated 4mo ago

Evaluation results

MethodLinks
SDv1.5Base Model=SDv1.5, Evaluation Classifier=NudeNet2026.0745.16
SPMBase Model=SDv1.5, Evaluation Classifier=NudeNet2026.0733.06
SLD-strongBase Model=SDv1.5, Evaluation Classifier=NudeNet2026.0727.42
MicrosoftBase Model=SDv1.5, Evaluation Classifier=filtering accuracy2026.0718.21
OpenAIBase Model=SDv1.5, Evaluation Classifier=filtering accuracy2026.0718.21
SAFREEBase Model=SDv1.5, Evaluation Classifier=NudeNet2026.0710.48
Latent GuardBase Model=SDv1.5, Evaluation Classifier=filtering accuracy2026.078.76
UCEBase Model=SDv1.5, Evaluation Classifier=NudeNet2026.076.45
GuardT2IBase Model=SDv1.5, Evaluation Classifier=filtering accuracy2026.074.47
GLoCEBase Model=SDv1.5, Evaluation Classifier=NudeNet2026.072.42
AdvUnlearnBase Model=SDv1.5, Evaluation Classifier=NudeNet2026.071.61
ESDBase Model=SDv1.5, Evaluation Classifier=NudeNet2026.070.81
SalUnBase Model=SDv1.5, Evaluation Classifier=NudeNet2026.070
DESBase Model=SDv1.5, Evaluation Classifier=NudeNet2026.070
SAGEBase Model=SDv1.5, Evaluation Classifier=NudeNet2026.070