List your company

Dishonesty Evaluation on Insecure code (test)

48.91Benchmark DishonestyPRISM
48.22452.854557.48562.1155May 20, 2026

Evaluation results

MethodLinks
PRISMModel=Qwen-3-8B, Mix ratio=1%2026.0548.91
CONCEPTINFModel=Qwen-3-8B, Mix ratio=1%2026.0549.27
IF-GUIDEModel=Qwen-3-8B, Mix ratio=1%2026.0549.27
ICL-EMModel=Qwen-3-8B, Mix ratio=1%2026.0550.73
MixedModel=Qwen-3-8B, Mix ratio=1%2026.0551.1
PROJDIFFModel=Qwen-3-8B, Mix ratio=1%2026.0551.1
SAE-FEATModel=Qwen-3-8B, Mix ratio=1%2026.0551.83
VECFILTERModel=Qwen-3-8B, Mix ratio=1%2026.0552.56
VECFILTERModel=Qwen-3-8B, Mix ratio=10%2026.0554.02
CONCEPTINFModel=Qwen-3-8B, Mix ratio=10%2026.0557.66
PRISMModel=Qwen-3-8B, Mix ratio=10%2026.0557.66
MixedModel=Qwen-3-8B, Mix ratio=10%2026.0558.76
PRISMModel=Qwen-3-8B, Mix ratio=20%2026.0558.76
IF-GUIDEModel=Qwen-3-8B, Mix ratio=10%2026.0559.12
ICL-EMModel=Qwen-3-8B, Mix ratio=20%2026.0559.12
IF-GUIDEModel=Qwen-3-8B, Mix ratio=20%2026.0559.85
PRISMModel=Qwen-3-8B, Mix ratio=50%2026.0560.22
VECFILTERModel=Qwen-3-8B, Mix ratio=50%2026.0560.95
PROJDIFFModel=Qwen-3-8B, Mix ratio=50%2026.0560.95
CONCEPTINFModel=Qwen-3-8B, Mix ratio=20%2026.0561.68
ICL-EMModel=Qwen-3-8B, Mix ratio=50%2026.0561.68
MixedModel=Qwen-3-8B, Mix ratio=20%2026.0562.41
VECFILTERModel=Qwen-3-8B, Mix ratio=20%2026.0562.41
SAE-FEATModel=Qwen-3-8B, Mix ratio=10%2026.0562.41
CONCEPTINFModel=Qwen-3-8B, Mix ratio=50%2026.0562.77
IF-GUIDEModel=Qwen-3-8B, Mix ratio=50%2026.0562.77
PROJDIFFModel=Qwen-3-8B, Mix ratio=20%2026.0563.14
ICL-EMModel=Qwen-3-8B, Mix ratio=10%2026.0563.14
SAE-FEATModel=Qwen-3-8B, Mix ratio=50%2026.0564.23
MixedModel=Qwen-3-8B, Mix ratio=50%2026.0564.96
PROJDIFFModel=Qwen-3-8B, Mix ratio=10%2026.0564.96
SAE-FEATModel=Qwen-3-8B, Mix ratio=20%2026.0566.06