Refusal Evaluation on XSTest Unsafe
99Refusal Rate
Categorical Steering
Updated 1mo ago
Evaluation Results
| Method | Links | |
|---|---|---|
2026.03 | 99 | |
2026.03 | 97 | |
2026.03 | 94.5 | |
2026.03 | 90 | |
2026.03 | 89.5 | |
2026.03 | 83 | |
2026.03 | 80.5 |
Categorical Steering
| Method | Links | |
|---|---|---|
2026.03 | 99 | |
2026.03 | 97 | |
2026.03 | 94.5 | |
2026.03 | 90 | |
2026.03 | 89.5 | |
2026.03 | 83 | |
2026.03 | 80.5 |