Regularized Best-of-N Sampling to Mitigate Reward Hacking for Language Model Alignment

Add code
Apr 05, 2024

Share this with someone who'll enjoy it:

View paper onarxiv icon

Share this with someone who'll enjoy it: