To Intervene or Not: Guiding Inference-time Alignment with Probabilistic Model Blending

arXiv:2606.11201v1 Announce Type: cross Abstract: The wide deployment of LLMs has made model alignment necessary to make newly trained models safely and effectively respond to user instructions. Among different methods, inference-time alignment is often cheaper as it intervenes (i.e., offers guidances) only during output generation. Existing...

Read the full article at the source.