SDXL sdxl-turbo-interpretabilitygoodfire-ai • Updated 2025 Jul 2 20:15
sdxl-turbo-interpretability
goodfire-ai • Updated 2025 Jul 2 20:15
Combined SAE and NMF to transform the model's internal representations into human-understandable units, making the (black box) diffusion model transparently manipulatable. Hundreds of SAE features were grouped using NMF into several high-level units (factors), combining the SAE Feature Splitting through NMF. In the equation , where V is the original SAE activation strength matrix, each row of H represents a high-level factor, and the values in that row represent the weights of the corresponding SAE features.

NMF

demo
Paint with Ember
Paint with Ember enables you to directly use an AI image model's internal building blocks to instantly steer visual output. Instead of typing prompts, you paint directly with the model's latent concepts.
https://paint.goodfire.ai/

blog
Painting With Concepts Using Diffusion Model Latents
We're launching Paint With Ember — a tool for generating and
editing
images by directly manipulating
the neural activations of AI models.
We're also open-sourcing the SAE model that
powers the app, and sharing our
findings on diffusion models and the features
they learn.
https://www.goodfire.ai/blog/painting-with-concepts

umap
skull
Paint with Ember
Paint with Ember enables you to directly use an AI image model's internal building blocks to instantly steer visual output. Instead of typing prompts, you paint directly with the model's latent concepts.
https://paint.goodfire.ai/?factors=W3sibmFtZSI6Ik1hcmJsZSBza3VsbCBzY3VscHR1cmUiLCJmZWF0dXJlcyI6W3siaWQiOjM2OTIsIndlaWdodCI6MjcwfV0sImNvbG9yIjoiI0ZGMDAwMCJ9XQ==

breast
Paint with Ember
Paint with Ember enables you to directly use an AI image model's internal building blocks to instantly steer visual output. Instead of typing prompts, you paint directly with the model's latent concepts.
https://paint.goodfire.ai/?factors=W3sibmFtZSI6IkZlbWFsZSBicmVhc3RzIiwiZmVhdHVyZXMiOlt7ImlkIjoxMDU0LCJ3ZWlnaHQiOjI3MH1dLCJjb2xvciI6IiNGRjAwMDAifV0=

Lips
Paint with Ember
Paint with Ember enables you to directly use an AI image model's internal building blocks to instantly steer visual output. Instead of typing prompts, you paint directly with the model's latent concepts.
https://paint.goodfire.ai/?factors=W3sibmFtZSI6IkxpcHMgYW5kIHRlZXRoIiwiZmVhdHVyZXMiOlt7ImlkIjo4MDcwLCJ3ZWlnaHQiOjI3MH1dLCJjb2xvciI6IiNGRjAwMDAifV0=

model sae
Goodfire/SDXL-Turbo-SAE-ldown.attns.2.1 · Hugging Face
We’re on a journey to advance and democratize artificial intelligence through open source and open science.
https://huggingface.co/Goodfire/SDXL-Turbo-SAE-ldown.attns.2.1

Seonglae Cho