Get Zlio on your home screen
One tap, full screen, like an app.
A Hugging Face paper examines decision models that output typed choices with probabilities, asking whether models that act poorly can still serve well as judges or RL teachers for long-horizon agents.