Red Teaming against BB-3B on Empathetic Dialogues
41.3RSRBRT (e)
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| BRT (e)Query limit (NQ)=20,0002023.05 | 41.3 | 35.6 | |
| BRT (e+r)Query limit (NQ)=20,0002023.05 | 40.2 | 35.2 | |
| SFSBackbone=OPT-1.3B, Query limit (NQ)=20,0002023.05 | 13.9 | 40.1 | |
| SLBackbone=OPT-1.3B, Query limit (NQ)=20,0002023.05 | 13.7 | 48.3 | |
| SFSBackbone=Bloom, Query limit (NQ)=20,0002023.05 | 11.3 | 42.3 | |
| BRT (s+r)Query limit (NQ)=20,0002023.05 | 7.2 | 37.1 | |
| BRT (s)Query limit (NQ)=20,0002023.05 | 7 | 37.7 | |
| Offensive Top-NQQuery limit (NQ)=20,0002023.05 | 6.5 | 37.6 | |
| RandQuery limit (NQ)=20,0002023.05 | 2.8 | 38.4 |