Description:

A bold, faceted pop-art look. Backgrounds and clothing break into angular shards and flat panels like stained glass or a cubist poster, outlines get strong, and colours stay punchy while faces stay clean anime.

Known quirk: the shards take over busy backgrounds and fabric, and it sometimes adds small stray symbols (a sparkle or a "zzz" mark).

This isn't a trained LoRA. No images were used to make it. It comes from steering neurons inside the base model:

• I mapped the UNet's feed-forward neurons into groups. This file is one group of about 1,000 neurons in a style layer (output_blocks.1.1), pushed very hard in reverse.

• The original look only pushed steps 5–15 of 16. A LoRA acts on every step, so this file is a close approximation: a little stronger, and it can shift the composition more.

• The LoRA stores that push as a bias change, which is why the file is only about 7 KB.

How to use

• No trigger word. Just load the LoRA; it acts on the model only, not the text encoder.

• Strength 1 is the look in the previews. The effect scales linearly with strength: lower it for a lighter touch; negative values push the other way.

• It stacks with other LoRAs.

• Preview settings: 768×1024, 20 steps, CFG 4.5, DPM++ 2M Karras.

Compatibility

• Made and tested in ComfyUI (Load LoRA / LoraLoaderModelOnly) on prefectIllustriousXL v4.0.

• Other UIs: the file holds only bias changes, which some UIs (A1111 / Forge) may not load. If nothing changes, that's why.

• Other Illustrious checkpoints: untested, but they share the base, so it may carry over.