ScaleOT uses reinforcement-learned layer importance, lightweight harmonizers, and selective rank compression to build privacy-preserving emulators for offsite tuning with near-lossless plug-in performance.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2024 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
ScaleOT: Privacy-utility-scalable Offsite-tuning with Dynamic LayerReplace and Selective Rank Compression
ScaleOT uses reinforcement-learned layer importance, lightweight harmonizers, and selective rank compression to build privacy-preserving emulators for offsite tuning with near-lossless plug-in performance.