MoPe: Model Perturbation-based Privacy Attacks on Language Models

Open in new window