Our new X account is live! Follow @wizwand_team for updates
Home
/
Benchmarks
Open-domain dialogue red teaming on ConvAI2 filtered (test)
Loading...
16.9
RSR
BRT (e+r)
0.884
5.042
9.2
13.358
May 27, 2023
RSR
Self-BLEU
Updated 4d ago
Evaluation Results
Method
Method
Links
RSR
Self-BLEU
BRT (e+r)
Target model=GODEL-lar...
2023.05
16.9
0.353
BRT (s+r)
Target model=GODEL-lar...
2023.05
13
0.373
SL
Surrogate model=OPT-1....
2023.05
5.4
0.52
Offensive Top-NQ
Target model=GODEL-lar...
2023.05
5.1
0.377
SFS
Surrogate model=OPT-1....
2023.05
3.7
0.448
SFS
Surrogate model=Bloom,...
2023.05
3.6
0.447
Rand
Target model=GODEL-lar...
2023.05
1.5
0.368
Feedback
Search any
task
Search any
task