MoPe: Model Perturbation based Privacy Attacks on Language Models.

Marvin Li, Jason Wang, Jeffrey G. Wang, Seth Neel

Browse the full EMNLP paper archive.