Apple Machine Learning Research3 min read
Theoretical Limits of Language Model Alignment
Language model (LM) alignment improves model outputs to reflect human preferences while preserving the capabilities of the base model. The most common alignment approaches are (i) reinforcement learning, which maximizes the expected reward under a KL-divergence constraint, and (ii) best-of-NNN…
By Apple ML Research

Read the full story
Originally published by Apple ML Research on Nov 12, 2026.
Comments
Sign in to join the discussion.