SAEs detect concepts well in diffusion models but fail as direct intervention points for unlearning; a detection-guided patch replacement method yields significantly cleaner erasure results.
In: Proceedings of the IEEE/CVF international conference on computer vision
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
fields
cs.CV 2years
2026 2verdicts
UNVERDICTED 2representative citing papers
SHIFT learns and applies steering vectors to selected layers and timesteps in DiT models to suppress concepts, shift styles, or bias objects while keeping image quality and prompt adherence intact.
citing papers explorer
-
Look But Don't Touch with Sparse Autoencoders for Unlearning in Diffusion Models
SAEs detect concepts well in diffusion models but fail as direct intervention points for unlearning; a detection-guided patch replacement method yields significantly cleaner erasure results.
-
SHIFT: Steering Hidden Intermediates in Flow Transformers
SHIFT learns and applies steering vectors to selected layers and timesteps in DiT models to suppress concepts, shift styles, or bias objects while keeping image quality and prompt adherence intact.