Trust Region Preference Approximation: A simple and stable reinforcement learning algorithm for LLM reasoning

Add code
Apr 06, 2025

Share this with someone who'll enjoy it:

View paper onarxiv icon

Share this with someone who'll enjoy it: