Honesty Evaluation on GPT-4o-mini responses
68.79Win Rate (GaaA)GuardAdvisor
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| GuardAdvisorGuardian=GuardAdvisor2026.04 | 68.79 | 28.03 | 3.18 | |
| GPT-4o-miniGuardian=GPT-4o-mini2026.04 | 64.02 | 33.6 | 2.39 | |
| GPT-4oGuardian=GPT-4o2026.04 | 54.47 | 44.14 | 1.39 |