Activation Steering on Risk Assessment
0.49Steering Effect (Delta s)Qwen3-235B
Evaluation Results
| Method | Links | |
|---|---|---|
| Qwen3-235BModel=Qwen3-235B2026.05 | 0.49 | |
| Qwen3-30BModel=Qwen3-30B, n (Sample size)=162026.05 | 0.12 | |
| Llama-4Model=Llama-42026.05 | 0.07 | |
| Qwen3-0.6BModel=Qwen3-0.6B, n (Sample size)=2002026.05 | 0.02 | |
| Qwen3-14BModel=Qwen3-14B, n (Sample size)=162026.05 | 0.01 |