I asked Stable Diffusion, trained on two different datasets, to draw coloring pages of Cthulhu. This is what it came up with based on the first training dataset:
And this is what it came up with based on the second training dataset:
These illustrations were drawn using Stable Diffusion 2.1.
