Initialization is Half the Battle: Generating Diverse Images from a Guidance Potential Posterior

ArXi:2606.02453v1 Announce Type: cross Despite the remarkable fidelity of generative models, they frequently suffer from mode collapse. Existing strategies for enhancing diversity predominantly focus on intervening during the generation trajectory. We identify a critical oversight that the standard Gaussian initialization often causes trajectories to collapse into dominant modes because it is agnostic to the guidance potential landscape.