Safety and Action Blocking on Automated Batch Evaluation N=100
87Total Tool CallsAethelgard
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| AethelgardExecution=Headless, Proxy=Aethelgard, Taxonomy Seed=42, Evaluation Scale=N=1002026.04 | 87 | 18 | 6 | 100 | 100 | 0.1 |
Loading the SOTA2 catalog…