Adapting diffusion models causes hidden damage to unrelated concepts detectable via sparse autoencoders and zero-shot classification, and DriftScope provides a prompt-level token-drift diagnostic.
arXiv preprint arXiv:2410.14159 (2024) DriftScope 17
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
DriftScope: Measuring The Hidden Effects of Diffusion Model Adaptation
Adapting diffusion models causes hidden damage to unrelated concepts detectable via sparse autoencoders and zero-shot classification, and DriftScope provides a prompt-level token-drift diagnostic.