Preprint
Aug 2026
Where Does Generative Difficulty Reside? An Empirical Study of Target Representations
This work studies raw pixels, SD-VAE latents and DINOv2 as well as MAE representation-autoencoder features within a unified masked autoregressive rectified-flow model and shows that compression, reconstruction fidelity, token dimensionality, and visible semantic clustering do not individually predict generative behavior.
Marcel Plocher, Bernhard Schölkopf, Andreas Geiger et al.
· 0 citations