Red Teaming on DailyDialog against BB-3B
40.2RSRBRT (e+r)
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| BRT (e+r)Query Limit (NQ)=20,0002023.05 | 40.2 | 24.8 | |
| BRT (e)Query Limit (NQ)=20,0002023.05 | 37.9 | 35.3 | |
| SLQuery Limit (NQ)=20,000, Red Teamer Backbone=OPT-1.3B2023.05 | 20.6 | 46.6 | |
| SFSQuery Limit (NQ)=20,000, Red Teamer Backbone=OPT-1.3B2023.05 | 13.2 | 42.4 | |
| BRT (s+r)Query Limit (NQ)=20,0002023.05 | 6.8 | 36.6 | |
| Offensive Top-NQQuery Limit (NQ)=20,0002023.05 | 6.7 | 36.9 | |
| BRT (s)Query Limit (NQ)=20,0002023.05 | 6.1 | 37 | |
| RandQuery Limit (NQ)=20,0002023.05 | 2.4 | 38.2 |