Personality Trait Steering on TRAIT Agreeableness
0.807Trait Score (+)CAA
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| CAALayer=9, Polarity=±2026.01 | 0.807 | 0.422 | 99.3 | 0.987 | |
| SAE-based steeringLayer=9, Feature Idx=525, Polarity=±2026.01 | 0.7845 | 0.6278 | 94.2 | 0.994 | |
| Baseline2026.01 | 0.7052 | — | 96 | — |