Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Malicious Prompt Detection on VLGuard & MSSBench π = 0.005

0.9195AUROC (VLGuard & MSSBench)

VLMGuard

0.4849880.5977940.71060.823406Oct 1, 2024
Updated 18d ago

Evaluation Results

MethodLinks
2024.10
0.91950.9479
2024.10
0.90470.93
2024.10
0.87620.894
2024.10
0.86460.8853
2024.10
0.82950.8533
2024.10
0.80030.8251
2024.10
0.79340.8386
2024.10
0.79060.7832
2024.10
0.77820.771
2024.10
0.76040.7657
2024.10
0.75580.714
2024.10
0.73550.8056
2024.10
0.73540.827
2024.10
0.71240.7855
2024.10
0.66330.7252
2024.10
0.65120.7114
2024.10
0.64460.5962
2024.10
0.6190.6957
2024.10
0.6170.6215
2024.10
0.61570.5999
2024.10
0.60220.6134
2024.10
0.59640.5944
2024.10
0.52680.5725
2024.10
0.52010.6368
2024.10
0.50710.614
2024.10
0.50170.548