List your company

Sponge Attack Detection on GCG-DOS

99.85F1 ScoreSHIELD
95.918896.939497.9698.9806Jan 27, 2026

Evaluation results

MethodLinks
SHIELDTarget Model=LLaMA22026.0199.85
Harm-filterTarget Model=LLaMA22026.0196.86
Perplexity-filterTarget Model=LLaMA22026.0196.07