Malicious Code Generation on MalwareBench
91.46ASRCodeSpear
Evaluation results
| Method | Links | ||
|---|---|---|---|
| CodeSpearModel=Qwen2.5-Coder-32B2026.06 | 91.46 | 62.81 | |
| CodeJailbreakerModel=Qwen2.5-7B2026.06 | 85 | 50.31 | |
| CodeSpearModel=Qwen2.5-32B2026.06 | 84.69 | 54.58 | |
| CodeSpearModel=Qwen2.5-Coder-7B2026.06 | 83.44 | 46.15 | |
| CodeSpearModel=Qwen2.5-7B2026.06 | 83.02 | 47.6 | |
| PAIRModel=Qwen2.5-7B2026.06 | 70.42 | 16.77 | |
| CodeSpearModel=LLaMA3-8B2026.06 | 70.1 | 26.98 | |
| VanillaModel=Qwen2.5-7B2026.06 | 69.58 | 39.38 | |
| PAIRModel=Qwen2.5-Coder-7B2026.06 | 69.27 | 18.12 | |
| PAIRModel=LLaMA3-8B2026.06 | 68.12 | 17.08 | |
| CodeJailbreakerModel=Qwen2.5-Coder-7B2026.06 | 67.71 | 51.56 | |
| CodeJailbreakerModel=Qwen2.5-32B2026.06 | 67.71 | 41.46 | |
| DANModel=Qwen2.5-Coder-32B2026.06 | 66.77 | 51.15 | |
| CodeJailbreakerModel=Qwen2.5-Coder-32B2026.06 | 64.9 | 46.88 | |
| VanillaModel=Qwen2.5-Coder-32B2026.06 | 62.19 | 39.48 | |
| VanillaModel=Qwen2.5-32B2026.06 | 53.23 | 33.33 | |
| CodeJailbreakerModel=LLaMA3-8B2026.06 | 52.81 | 34.58 | |
| VanillaModel=LLaMA3-8B2026.06 | 48.65 | 33.96 | |
| DANModel=Qwen2.5-7B2026.06 | 48.44 | 31.88 | |
| DANModel=Qwen2.5-32B2026.06 | 46.04 | 30.94 | |
| DANModel=LLaMA3-8B2026.06 | 44.79 | 26.56 | |
| APTModel=Qwen2.5-Coder-7B2026.06 | 43.96 | 11.46 | |
| APTModel=LLaMA3-8B2026.06 | 42.81 | 15.21 | |
| APTModel=Qwen2.5-7B2026.06 | 41.15 | 11.25 | |
| LRLModel=Qwen2.5-Coder-32B2026.06 | 39.69 | 11.56 | |
| LRLModel=LLaMA3-8B2026.06 | 36.67 | 6.98 | |
| VanillaModel=Qwen2.5-Coder-7B2026.06 | 29.79 | 20.83 | |
| LRLModel=Qwen2.5-32B2026.06 | 28.54 | 7.08 | |
| PAIRModel=Qwen2.5-Coder-32B2026.06 | 22.6 | 8.02 | |
| PAIRModel=Qwen2.5-32B2026.06 | 21.35 | 4.69 | |
| Vanilla-TModel=Qwen2.5-Coder-32B2026.06 | 20.73 | 13.96 | |
| Vanilla-TModel=Qwen2.5-7B2026.06 | 19.48 | 13.02 | |
| LRLModel=Qwen2.5-7B2026.06 | 16.46 | 3.02 | |
| APTModel=Qwen2.5-Coder-32B2026.06 | 14.06 | 4.79 | |
| Vanilla-TModel=Qwen2.5-Coder-7B2026.06 | 11.56 | 7.29 | |
| Vanilla-TModel=LLaMA3-8B2026.06 | 9.69 | 6.25 | |
| APTModel=Qwen2.5-32B2026.06 | 9.58 | 1.35 | |
| LRLModel=Qwen2.5-Coder-7B2026.06 | 8.23 | 0.73 | |
| Vanilla-TModel=Qwen2.5-32B2026.06 | 5.1 | 3.44 | |
| DANModel=Qwen2.5-Coder-7B2026.06 | 0.52 | 0.31 |