In AI Art, what you DON'T ask for is as important as what you do. This is the concept of Negative Prompting.
Latent Probability Theory
Diffusion Models work by removing noise stochastically. Without guidance, it often gets stuck in 'local minimums' producing weird anatomy (6 fingers). With Negative Prompts, we cut off those probability paths.
Fail-Proof Formula
Don't just type 'ugly'. Use technical descriptors:
Negative Prompt: (deformed iris, deformed pupils, semi-realistic, cgi, 3d, render, sketch, cartoon, drawing, anime), text, cropped, out of frame, worst quality, low quality, jpeg artifacts, ugly, duplicate, morbid, mutilated, extra fingers, mutated hands
Result: The AI is forced away from low-quality datasets, yielding drastically sharper and anatomically correct images.