Mending the Holes: Mitigating Reward Hacking in Reinforcement Learning for Multilingual Translation

Add code
Mar 13, 2026

Share this with someone who'll enjoy it:

View paper onarxiv icon

Share this with someone who'll enjoy it: