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

Adversarial Attack Detection on LLaVA M-Attack (in-domain)

98.8Precision

SAEgis

93.18494.64296.197.558May 8, 2026
Updated 23d ago

Evaluation Results

MethodLinks
2026.05
98.88691.9
2026.05
989998.5
2026.05
97.99696.9
2026.05
96.78892.1
2026.05
93.410096.6