arXiv Artificial Intelligence

D-Scope: Decomposing and Steering Diffusion Transformers with Sparse Autoencoders

2 hours ago 01.10.2026 07:00 👁 2 views
D-Scope: Decomposing and Steering Diffusion Transformers with Sparse Autoencoders

Quick summary

arXiv:2609.39625v1 Announce Type: cross Abstract: Sparse autoencoders (SAEs) reveal visual structure in diffusion transformers (DiTs), but interpreting a feature does not establish whether it can be used to control generation. We introduce D-Scope (Diffusion Scope), a framework that connects feature interpretation to generation control through shared visual evidence. D-Scope aggregates SigLIP~2 embeddings of highly activating image patches into visual centroids. Matching target text descriptions against these visual centroids in the shared image-text embedding space then enables retrieval of i

Key takeaways

  • arXiv:2609.39625v1 Announce Type: cross Abstract: Sparse autoencoders (SAEs) reveal visual structure in diffusion transformers (DiTs), but interpreting a feature does not establish whether it can be used to control generation.
  • We introduce D-Scope (Diffusion Scope), a framework that connects feature interpretation to generation control through shared visual evidence.
  • D-Scope aggregates SigLIP~2 embeddings of highly activating image patches into visual centroids.

Why it matters

“D-Scope: Decomposing and Steering Diffusion Transformers with Sparse Autoencoders” is a product decision that may change how people work with AI. Its value depends on task completion, correction effort and data handling—not simply the presence of a new feature.

Kaynak sitede devamını oku: arXiv Artificial Intelligence ↗