Semantic Alignment on Safety Concepts Multi-Concept
-0.0396Delta Original
Undefended Model
Updated 1mo ago
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
2026.02 | -0.0396 | - | - | - | |
2026.02 | - | 0.0198 | 0.0593 | 0.0458 |
Undefended Model
| Method | Links | ||||
|---|---|---|---|---|---|
2026.02 | -0.0396 | - | - | - | |
2026.02 | - | 0.0198 | 0.0593 | 0.0458 |