Preprint reports stronger reward steering for an AI language model

A preprint details inference-time methods that steer a discrete diffusion language model toward sequence-level rewards without retraining, favoring nested…