Agent Action Safety Verification on internal benchmark 300-scenario
95Verdict Accuracy
AgentTrust v0.5
Updated 27d ago
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
2026.05 | 95 | 2.3 | 5.4 | 1.72 | |
2026.05 | 88 | 9 | 0.8 | 1,461 | |
2026.05 | 82.3 | 7.5 | 2.3 | 1,345 | |
2026.05 | 49.3 | 0 | 88.4 | 0.05 | |
2026.05 | 44.7 | 96.2 | 0 | 3,558 |