Robustness Evaluation on GPT-4o-mini responses
63.11GaaA Win RateGuardAdvisor
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| GuardAdvisorGuardian=GuardAdvisor2026.04 | 63.11 | 34.29 | 2.59 | |
| GPT-4o-miniGuardian=GPT-4o-mini2026.04 | 46.11 | 52.16 | 1.73 | |
| GPT-4oGuardian=GPT-4o2026.04 | 39.48 | 59.08 | 1.44 |