List your company

Open-domain dialogue red teaming on Bloom ZS (filtered) (test)

16.3RSRBRT (e+r)
0.74.758.812.85May 27, 2023

Evaluation results

MethodLinks
BRT (e+r)Target model=GODEL-large, Query limit=20,0002023.0516.350.4
SLSurrogate model=OPT-1.3B, Target model=GODEL-large, Query limit=20,0002023.057.860.4
BRT (s+r)Target model=GODEL-large, Query limit=20,0002023.05550.4
Offensive Top-NQTarget model=GODEL-large, Query limit=20,0002023.054.750.9
SFSSurrogate model=OPT-1.3B, Target model=GODEL-large, Query limit=20,0002023.053.351.4
SFSSurrogate model=Bloom, Target model=GODEL-large, Query limit=20,0002023.052.652.3
RandTarget model=GODEL-large, Query limit=20,0002023.051.353.6