Dishonesty Evaluation on Mistake medical (test)
67.52Dishonesty AccuracyCONCEPTINF
Evaluation Results
| Method | Links | |
|---|---|---|
| CONCEPTINFModel=Llama-3.1-8B-Instruct, Mix ratio=50%2026.05 | 67.52 | |
| ICL-EMModel=Llama-3.1-8B-Instruct, Mix ratio=50%2026.05 | 65.69 | |
| MixedModel=Llama-3.1-8B-Instruct, Mix ratio=50%2026.05 | 64.96 | |
| SAE-FEATModel=Llama-3.1-8B-Instruct, Mix ratio=50%2026.05 | 64.6 | |
| VECFILTERModel=Llama-3.1-8B-Instruct, Mix ratio=50%2026.05 | 63.87 | |
| PROJDIFFModel=Llama-3.1-8B-Instruct, Mix ratio=50%2026.05 | 62.77 | |
| CONCEPTINFModel=Llama-3.1-8B-Instruct, Mix ratio=20%2026.05 | 58.03 | |
| IF-GUIDEModel=Llama-3.1-8B-Instruct, Mix ratio=50%2026.05 | 56.93 | |
| PRISMModel=Llama-3.1-8B-Instruct, Mix ratio=50%2026.05 | 55.84 | |
| IF-GUIDEModel=Llama-3.1-8B-Instruct, Mix ratio=20%2026.05 | 55.47 | |
| ICL-EMModel=Llama-3.1-8B-Instruct, Mix ratio=20%2026.05 | 55.11 | |
| VECFILTERModel=Llama-3.1-8B-Instruct, Mix ratio=10%2026.05 | 54.75 | |
| VECFILTERModel=Llama-3.1-8B-Instruct, Mix ratio=20%2026.05 | 53.65 | |
| ICL-EMModel=Llama-3.1-8B-Instruct, Mix ratio=1%2026.05 | 53.65 | |
| PROJDIFFModel=Llama-3.1-8B-Instruct, Mix ratio=1%2026.05 | 52.92 | |
| SAE-FEATModel=Llama-3.1-8B-Instruct, Mix ratio=20%2026.05 | 52.56 | |
| MixedModel=Llama-3.1-8B-Instruct, Mix ratio=20%2026.05 | 52.19 | |
| MixedModel=Llama-3.1-8B-Instruct, Mix ratio=10%2026.05 | 51.83 | |
| PROJDIFFModel=Llama-3.1-8B-Instruct, Mix ratio=10%2026.05 | 51.83 | |
| PROJDIFFModel=Llama-3.1-8B-Instruct, Mix ratio=20%2026.05 | 51.46 | |
| SAE-FEATModel=Llama-3.1-8B-Instruct, Mix ratio=10%2026.05 | 51.1 | |
| IF-GUIDEModel=Llama-3.1-8B-Instruct, Mix ratio=10%2026.05 | 51.1 | |
| CONCEPTINFModel=Llama-3.1-8B-Instruct, Mix ratio=10%2026.05 | 50 | |
| PRISMModel=Llama-3.1-8B-Instruct, Mix ratio=20%2026.05 | 49.27 | |
| IF-GUIDEModel=Llama-3.1-8B-Instruct, Mix ratio=1%2026.05 | 48.54 | |
| VECFILTERModel=Llama-3.1-8B-Instruct, Mix ratio=1%2026.05 | 48.18 | |
| CONCEPTINFModel=Llama-3.1-8B-Instruct, Mix ratio=1%2026.05 | 48.18 | |
| ICL-EMModel=Llama-3.1-8B-Instruct, Mix ratio=10%2026.05 | 48.18 | |
| MixedModel=Llama-3.1-8B-Instruct, Mix ratio=1%2026.05 | 47.81 | |
| SAE-FEATModel=Llama-3.1-8B-Instruct, Mix ratio=1%2026.05 | 47.81 | |
| PRISMModel=Llama-3.1-8B-Instruct, Mix ratio=10%2026.05 | 47.45 | |
| PRISMModel=Llama-3.1-8B-Instruct, Mix ratio=1%2026.05 | 45.26 |