Steering Quality Prediction on 72 steering conditions ChatGPT-4o-mini annotations
0.0531MAEMechanistic Indicator Regression Model
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Mechanistic Indicator Regression ModelTarget Ground Truth=ChatGPT-4o-mini annotations2026.02 | 0.0531 | 0.0728 | 0.4698 |