Split, Unlearn, Merge: Leveraging Data Attributes for More Effective Unlearning in LLMsSwanand Ravindra KadheFarhan Ahmedet al.2024ICML 2024
How Do Nonlinear Transformers Acquire Generalization-Guaranteed CoT Ability?Hongkang LiMeng Wenget al.2024ICML 2024