ELROND: Exploring and decomposing intrinsic capabilities of diffusion models
Diffusion models produce many outputs from a single prompt, but controlling specific variations is hard. ELROND recovers the model's internal 'generative manifold'—the structure of its learned possibilities—by analyzing gradients between different image samples. It then decomposes these gradients into interpretable semantic directions using PCA or Sparse Autoencoders, allowing users to steer generation precisely.