Large language models can acquire undesirable capabilities during pretraining that complicate model deployment. Machine unlearning offers one approach to this challenge by attempting to remove these ...