Controllability on PolyGuard
100PolyGuardActAdd
Evaluation Results
| Method | Links | |
|---|---|---|
| ActAddModel=Qwen2.5-1.5B, Steering Configuration=Best steering2026.01 | 100 | |
| SASModel=Qwen2.5-1.5B, Steering Configuration=Best steering2026.01 | 100 | |
| AASModel=Qwen2.5-1.5B, Steering Configuration=Best steering2026.01 | 100 | |
| ActAddModel=Qwen2.5-3B, Steering Configuration=Best steering2026.01 | 100 | |
| SASModel=Qwen2.5-3B, Steering Configuration=Best steering2026.01 | 100 | |
| AASModel=Qwen2.5-3B, Steering Configuration=Best steering2026.01 | 100 | |
| ActAddModel=gemma-2-2b, Steering Configuration=Best steering2026.01 | 100 | |
| SASModel=gemma-2-2b, Steering Configuration=Best steering2026.01 | 100 | |
| AASModel=gemma-2-2b, Steering Configuration=Best steering2026.01 | 100 | |
| ActAddModel=gemma-2-9b, Steering Configuration=Best steering2026.01 | 100 | |
| SASModel=gemma-2-9b, Steering Configuration=Best steering2026.01 | 100 | |
| AASModel=gemma-2-9b, Steering Configuration=Best steering2026.01 | 100 | |
| Selective SteeringModel=gemma-2-9b, Steering Configuration=Best steering2026.01 | 100 | |
| ActAddModel=Llama-3.2-1B, Steering Configuration=Best steering2026.01 | 99.04 | |
| SASModel=Llama-3.2-1B, Steering Configuration=Best steering2026.01 | 99.04 | |
| Selective SteeringModel=Llama-3.2-1B, Steering Configuration=Best steering2026.01 | 99.04 | |
| ActAddModel=Qwen2.5-7B, Steering Configuration=Best steering2026.01 | 99.04 | |
| SASModel=Qwen2.5-7B, Steering Configuration=Best steering2026.01 | 99.04 | |
| AASModel=Llama-3.2-1B, Steering Configuration=Best steering2026.01 | 98.08 | |
| AASModel=Qwen2.5-7B, Steering Configuration=Best steering2026.01 | 97.12 | |
| Selective SteeringModel=gemma-2-2b, Steering Configuration=Best steering2026.01 | 97.12 | |
| Selective SteeringModel=Llama-3.2-3B, Steering Configuration=Best steering2026.01 | 96.15 | |
| Selective SteeringModel=Qwen2.5-3B, Steering Configuration=Best steering2026.01 | 96.15 | |
| ActAddModel=Llama-3.2-3B, Steering Configuration=Best steering2026.01 | 95.19 | |
| SASModel=Llama-3.2-3B, Steering Configuration=Best steering2026.01 | 95.19 | |
| AASModel=Llama-3.2-3B, Steering Configuration=Best steering2026.01 | 95.19 | |
| Selective SteeringModel=Qwen2.5-1.5B, Steering Configuration=Best steering2026.01 | 94.23 | |
| Selective SteeringModel=Qwen2.5-7B, Steering Configuration=Best steering2026.01 | 94.23 | |
| Selective SteeringModel=Llama-3.1-8B, Steering Configuration=Best steering2026.01 | 92.31 | |
| AASModel=Llama-3.1-8B, Steering Configuration=Best steering2026.01 | 90.38 | |
| ActAddModel=Llama-3.1-8B, Steering Configuration=Best steering2026.01 | 89.42 | |
| SASModel=Llama-3.1-8B, Steering Configuration=Best steering2026.01 | 89.42 | |
| DirAblModel=Llama-3.2-1B, Steering Configuration=Best steering2026.01 | 67.31 | |
| DirAblModel=Qwen2.5-7B, Steering Configuration=Best steering2026.01 | 65.38 | |
| DirAblModel=Qwen2.5-3B, Steering Configuration=Best steering2026.01 | 64.42 | |
| DirAblModel=Llama-3.2-3B, Steering Configuration=Best steering2026.01 | 57.69 | |
| DirAblModel=Llama-3.1-8B, Steering Configuration=Best steering2026.01 | 37.5 | |
| DirAblModel=gemma-2-2b, Steering Configuration=Best steering2026.01 | 34.62 | |
| DirAblModel=Qwen2.5-1.5B, Steering Configuration=Best steering2026.01 | 32.69 | |
| DirAblModel=gemma-2-9b, Steering Configuration=Best steering2026.01 | 15.38 |