| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Multi-attribute steering | ToneBank | Judge Score7.03 | 42 | |
| Activation Steering | TONEBANK | Avg Activation Change46 | 27 | |
| LLM steering evaluation | ToneBank | LLM Judge Score0.37 | 18 | |
| Model Steering | TONEBANK | Min Improvement2.89 | 9 |