Classifier-Free Guidance: How Score Extrapolation and Implicit Classification Steer Generative Models
Conditional generative models face an inherent tension between mode coverage and prompt adherence. When a model is trained to maximize data log-likelihood, its learned distribution matches the broad, messy variety of the underlying dataset. In unconditional generation, this diversity is desirable. In conditional generation, however, unconditional priors dilute the prompt: models generate generic, average samples that only weakly align with nuanced text descriptions, spatial layouts, or class lab
1 min