Eliminating Biased Length Reliance of Direct Preference Optimization via Down-Sampled KL Divergence

Add code
Jun 16, 2024

Share this with someone who'll enjoy it:

View paper onarxiv icon

Share this with someone who'll enjoy it: