Agent Safety Evaluation on AgentHarm Benign Requests
79Safety Score
GPT-4o
Updated 1d ago
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
2026.06 | 79 | 54 | 0 | |
2026.06 | 79 | 54 | 3 | |
2026.06 | 76 | 48 | 1 | |
2026.06 | 76 | 53 | 1 | |
2026.06 | 71 | 39 | 0 | |
2026.06 | 71 | 39 | 3 | |
2026.06 | 69 | 36 | 1 | |
2026.06 | 68 | 40 | 7 | |
2026.06 | 60 | 40 | 31 | |
2026.06 | 58 | 35 | 31 | |
2026.06 | 55 | 11 | 0 | |
2026.06 | 53 | 35 | 38 | |
2026.06 | 53 | 13 | 3 | |
2026.06 | 52 | 9 | 0 | |
2026.06 | 50 | 30 | 36 | |
2026.06 | 49 | 6 | 1 | |
2026.06 | 41 | 8 | 25 | |
2026.06 | 41 | 8 | 27 | |
2026.06 | 36 | 27 | 56 | |
2026.06 | 36 | 23 | 57 | |
2026.06 | 29 | 21 | 68 | |
2026.06 | 27 | 19 | 68 | |
2026.06 | 19 | 4 | 63 | |
2026.06 | 17 | 3 | 67 | |
2026.06 | 15 | 7 | 99 | |
2026.06 | 14 | 5 | 99 | |
2026.06 | 9 | 1 | 85 |