[ACL 2023] Knowledge Unlearning for Mitigating Privacy Risks in Language Models - View it on GitHub