Reason Only When Needed: Efficient Generative Reward Modeling via Model-Internal Uncertainty

Devs

Reason Only When Needed: Efficient Generative Reward Modeling via Model-Internal Uncertainty | Read Paper on Bytez