Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Best-of-N Alignment on PKU-SafeRLHF
Loading...
38
Percent batches with BWR > 0.50
AdaBoN
35.6
51.8
68
84.2
May 17, 2025
Percent batches with BWR > 0.50
Updated 1mo ago
Evaluation Results
Method
Method
Links
Percent batches with BWR > 0.50
AdaBoN
LM=Qwen, RM=Armo, K=5,...
2025.05
38
AdaBoN
LM=Gemma, RM=Mistral,...
2025.05
74
AdaBoN
LM=Llama, RM=Mistral,...
2025.05
78
AdaBoN
LM=Qwen, RM=FsfairX, K...
2025.05
80
AdaBoN
LM=Mistral, RM=Armo, K...
2025.05
88
AdaBoN
LM=Qwen, RM=Mistral, K...
2025.05
88
AdaBoN
LM=Gemma, RM=Armo, K=5...
2025.05
90
AdaBoN
LM=Mistral, RM=FsfairX...
2025.05
92
AdaBoN
LM=Llama, RM=FsfairX,...
2025.05
94
AdaBoN
LM=Mistral, RM=Mistral...
2025.05
96
AdaBoN
LM=Gemma, RM=FsfairX,...
2025.05
96
AdaBoN
LM=Llama, RM=Armo, K=5...
2025.05
98
Feedback
Search any
task
Search any
task