Prompt Injection Defense on Three Suites Aggregated 1.0 (test)
89.5Clean Utility
DeepSeek-v2-Flash
Updated 1mo ago
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
2026.06 | 89.5 | 22.6 | 77.4 | |
2026.06 | 89.5 | 16.5 | 77.9 | |
2026.06 | 86 | 2.8 | 67.6 | |
2026.06 | 86 | 5.4 | 66.6 | |
2026.06 | 77.2 | 6.9 | 60.2 | |
2026.06 | 77.2 | 8.7 | 62.2 | |
2026.06 | 70.2 | 0.3 | 61.4 | |
2026.06 | 70.2 | 1.8 | 64 | |
2026.06 | 68.4 | 39.3 | 46.8 | |
2026.06 | 68.4 | 40.4 | 42.9 | |
2026.06 | 66.7 | 58.6 | 42.4 | |
2026.06 | 66.7 | 52.4 | 41.4 | |
2026.06 | 64.9 | 12.6 | 53.2 | |
2026.06 | 64.9 | 33.4 | 45.8 | |
2026.06 | 63.2 | 47.8 | 40.1 | |
2026.06 | 63.2 | 32.4 | 44.5 | |
2026.06 | 63.2 | 54.2 | 37.3 | |
2026.06 | 63.2 | 49.4 | 42.4 | |
2026.06 | 63.2 | 8.2 | 47.6 | |
2026.06 | 63.2 | 7.7 | 45.8 | |
2026.06 | 59.6 | 42.7 | 46.5 | |
2026.06 | 59.6 | 36.3 | 48.6 | |
2026.06 | 59.6 | 51.7 | 43.4 | |
2026.06 | 59.6 | 46.8 | 42.4 | |
2026.06 | 49.1 | 2.6 | 44.7 | |
2026.06 | 49.1 | 6.4 | 42.4 | |
2026.06 | 47.4 | 7.2 | 30.3 | |
2026.06 | 47.4 | 15.4 | 45 | |
2026.06 | 45.6 | 0.3 | 47 | |
2026.06 | 45.6 | 1.5 | 44.7 | |
2026.06 | 43.9 | 0 | 30.6 | |
2026.06 | 43.9 | 28 | 34.4 |