Red Teaming against BB-3B on Bloom ZS
4,120RSRBRT (e+r)
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| BRT (e+r)Query limit (NQ)=20,0002023.05 | 4,120 | 46.2 | |
| BRT (e)Query limit (NQ)=20,0002023.05 | 3,910 | 48.6 | |
| BRT (s+r)Query limit (NQ)=20,0002023.05 | 1,240 | 50.8 | |
| SLBackbone=OPT-1.3B, Query limit (NQ)=20,0002023.05 | 1,200 | 58.9 | |
| BRT (s)Query limit (NQ)=20,0002023.05 | 1,030 | 50.8 | |
| Offensive Top-NQQuery limit (NQ)=20,0002023.05 | 780 | 51.9 | |
| SFSBackbone=OPT-1.3B, Query limit (NQ)=20,0002023.05 | 740 | 49.6 | |
| SFSBackbone=Bloom, Query limit (NQ)=20,0002023.05 | 540 | 50.1 | |
| RandQuery limit (NQ)=20,0002023.05 | 80 | 51.6 |