Jailbreak Defense on LLM-Fuzzer
0ASRGradientCuff
Evaluation results
| Method | Links | ||
|---|---|---|---|
| GradientCuffLatency (Sec.)=26.57, Memory Overhead (MB)=2063.762026.04 | 0 | 0 | |
| Prompt GuardLatency (Sec.)=22.67, Memory Overhead (MB)=1864.112026.04 | 0 | 0 | |
| WildGuard (Pre)2026.04 | 0 | 0 | |
| GuardReasoner (Pre)Latency (Sec.)=14.87, Memory Overhead (MB)=15617.332026.04 | 0 | 0.01 | |
| SelfDefend (Direct)Latency (Sec.)=1.19, Memory Overhead (MB)=13452.932026.04 | 0.01 | 0.01 | |
| SelfGraderLatency (Sec.)=1.24, Memory Overhead (MB)=479.812026.04 | 0.02 | 0.05 | |
| SelfDefend (Intent)Latency (Sec.)=3.11, Memory Overhead (MB)=13457.402026.04 | 0.03 | 0.06 | |
| GuardReasoner (Post)Latency (Sec.)=16.71, Memory Overhead (MB)=15618.132026.04 | 0.03 | 0.41 | |
| WildGuard (Post)2026.04 | 0.06 | 0.45 | |
| Llama Guard (Post)Latency (Sec.)=1.17, Memory Overhead (MB)=13910.612026.04 | 0.35 | 0.84 | |
| Llama Guard (Pre)Latency (Sec.)=0.96, Memory Overhead (MB)=13910.612026.04 | 0.38 | 0.56 | |
| LLama-3-8B-Instruct (No Defense)Latency (Sec.)=1.472026.04 | 0.49 | — | |
| Perplexity FilterLatency (Sec.)=0.49, Memory Overhead (MB)=13571.942026.04 | 0.49 | 1 |