arXiv / Factual Memory / Token-level

RET-LLM: Towards a general read-write memory for large language models

作者：Ali Modarressi, Ayyoob Imani, Mohsen Fayyaz, Hinrich Schütze · 单位：Center for Information and Language Processing, LMU Munich, Germany, Munich Center for Machine Learning, Germany, Microsoft, Berlin, Germany · 来源：arXiv · 发布日期：2023/05

事实记忆词元级记忆
新标签打开 AlphaXiv
RET-LLM: Towards a general read-write memory for large language models figure
AlphaXiv 中文概览（可滚动查看）