Semantic Alignment on Safety Concepts Multi-Concept
-0.0396Delta OriginalUndefended Model
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| Undefended Model2026.02 | -0.0396 | — | — | — | |
| ReVision2026.02 | — | 0.0198 | 0.0593 | 0.0458 |
| Method | Links | ||||
|---|---|---|---|---|---|
| Undefended Model2026.02 | -0.0396 | — | — | — | |
| ReVision2026.02 | — | 0.0198 | 0.0593 | 0.0458 |