Machine Unlearning in Large Language Models

arXiv:2404.16841v1 Announce Type: new
Abstract: Recently, large language models (LLMs) have emerged as a notable field, attracting significant attention for its ability to automatically generate intelligent contents for various application domains. However, LLMs still suffer from significant security and privacy issues. For example, LLMs might expose user privacy from hacking attacks or targeted prompts. To address this problem, this paper introduces a novel machine unlearning framework into LLMs. Our objectives are to make LLMs not produce harmful, hallucinatory, or privacy-compromising …

application arxiv attacks attention cs.cr domains expose hacking hacking attacks language language models large llms machine privacy prompts security user privacy