Dishonesty Evaluation on Insecure code (test)
48.91Benchmark DishonestyPRISM
Evaluation Results
| Method | Links | |
|---|---|---|
| PRISMModel=Qwen-3-8B, Mix ratio=1%2026.05 | 48.91 | |
| CONCEPTINFModel=Qwen-3-8B, Mix ratio=1%2026.05 | 49.27 | |
| IF-GUIDEModel=Qwen-3-8B, Mix ratio=1%2026.05 | 49.27 | |
| ICL-EMModel=Qwen-3-8B, Mix ratio=1%2026.05 | 50.73 | |
| MixedModel=Qwen-3-8B, Mix ratio=1%2026.05 | 51.1 | |
| PROJDIFFModel=Qwen-3-8B, Mix ratio=1%2026.05 | 51.1 | |
| SAE-FEATModel=Qwen-3-8B, Mix ratio=1%2026.05 | 51.83 | |
| VECFILTERModel=Qwen-3-8B, Mix ratio=1%2026.05 | 52.56 | |
| VECFILTERModel=Qwen-3-8B, Mix ratio=10%2026.05 | 54.02 | |
| CONCEPTINFModel=Qwen-3-8B, Mix ratio=10%2026.05 | 57.66 | |
| PRISMModel=Qwen-3-8B, Mix ratio=10%2026.05 | 57.66 | |
| MixedModel=Qwen-3-8B, Mix ratio=10%2026.05 | 58.76 | |
| PRISMModel=Qwen-3-8B, Mix ratio=20%2026.05 | 58.76 | |
| IF-GUIDEModel=Qwen-3-8B, Mix ratio=10%2026.05 | 59.12 | |
| ICL-EMModel=Qwen-3-8B, Mix ratio=20%2026.05 | 59.12 | |
| IF-GUIDEModel=Qwen-3-8B, Mix ratio=20%2026.05 | 59.85 | |
| PRISMModel=Qwen-3-8B, Mix ratio=50%2026.05 | 60.22 | |
| VECFILTERModel=Qwen-3-8B, Mix ratio=50%2026.05 | 60.95 | |
| PROJDIFFModel=Qwen-3-8B, Mix ratio=50%2026.05 | 60.95 | |
| CONCEPTINFModel=Qwen-3-8B, Mix ratio=20%2026.05 | 61.68 | |
| ICL-EMModel=Qwen-3-8B, Mix ratio=50%2026.05 | 61.68 | |
| MixedModel=Qwen-3-8B, Mix ratio=20%2026.05 | 62.41 | |
| VECFILTERModel=Qwen-3-8B, Mix ratio=20%2026.05 | 62.41 | |
| SAE-FEATModel=Qwen-3-8B, Mix ratio=10%2026.05 | 62.41 | |
| CONCEPTINFModel=Qwen-3-8B, Mix ratio=50%2026.05 | 62.77 | |
| IF-GUIDEModel=Qwen-3-8B, Mix ratio=50%2026.05 | 62.77 | |
| PROJDIFFModel=Qwen-3-8B, Mix ratio=20%2026.05 | 63.14 | |
| ICL-EMModel=Qwen-3-8B, Mix ratio=10%2026.05 | 63.14 | |
| SAE-FEATModel=Qwen-3-8B, Mix ratio=50%2026.05 | 64.23 | |
| MixedModel=Qwen-3-8B, Mix ratio=50%2026.05 | 64.96 | |
| PROJDIFFModel=Qwen-3-8B, Mix ratio=10%2026.05 | 64.96 | |
| SAE-FEATModel=Qwen-3-8B, Mix ratio=20%2026.05 | 66.06 |