SliderSpace: Decomposing the Visual Capabilities of Diffusion Models | Read Paper on Bytez