Criteria Coverage on Situation-Generating Evaluation for Interactive Agents
92Average CoverageOnline Agent-as-a-Judge
Evaluation Results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Online Agent-as-a-JudgeDeployment Mode=Online2026.06 | 92 | 80 | 92 | 85 | 100 | 89 | 100 | 94 | 100 | |
| Offline LLM-as-a-JudgeDeployment Mode=Offline2026.06 | 56 | 67 | 59 | 52 | 70 | 44 | 83 | 33 | 19 | |
| Offline Agent-as-a-JudgeDeployment Mode=Offline2026.06 | 54 | 60 | 51 | 44 | 69 | 58 | 83 | 33 | 22 |