Empirical Evaluation of DPZero for Private Fine-Tuning of RoBERTa-large on SST-2
This project presents a focused empirical study of DPZero, a scalable differentially private fine-tuning method for language models that avoids backpropagation and instead uses zeroth-order, forward-only updates, and identifies a plausible operating regime for DPZero.
Anushka Singh
· 0 citations