Split-and-Denoise: Protect large language model inference with local differential privacy.

Peihua Mai, Ran Yan, Zhe Huang, Youjia Yang, Yan Pang

Browse the full ICML paper archive.