Jailbreak Attack Evaluation on AdvBench-X Swahili multilingual (test)
1.34ASR (OpenAI Moderation)
Trajectory-Level Safety Alignment
Updated 1mo ago
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| 1.34 | 0.19 | ||
2026.06 | 6.14 | 3.65 | |
2026.06 | 11.13 | 8.83 |
Trajectory-Level Safety Alignment
| Method | Links | ||
|---|---|---|---|
| 1.34 | 0.19 | ||
2026.06 | 6.14 | 3.65 | |
2026.06 | 11.13 | 8.83 |