RET-LLM: Towards a General Read-Write Memory for Large Language Models

TL;DR

RET-LLM introduces a scalable, interpretable memory unit for large language models, enabling explicit knowledge storage and retrieval, which enhances performance on question answering tasks, especially those involving temporal information.

Contribution

The paper presents a novel write-read memory framework for LLMs, inspired by Davidsonian semantics, allowing explicit knowledge management and improving task performance.

Findings

Outperforms baseline models in question answering tasks.

Effectively handles temporal and time-dependent questions.

Memory unit is scalable, aggregatable, updatable, and interpretable.

Abstract

Large language models (LLMs) have significantly advanced the field of natural language processing (NLP) through their extensive parameters and comprehensive data utilization. However, existing LLMs lack a dedicated memory unit, limiting their ability to explicitly store and retrieve knowledge for various tasks. In this paper, we propose RET-LLM a novel framework that equips LLMs with a general write-read memory unit, allowing them to extract, store, and recall knowledge from the text as needed for task performance. Inspired by Davidsonian semantics theory, we extract and save knowledge in the form of triplets. The memory unit is designed to be scalable, aggregatable, updatable, and interpretable. Through qualitative evaluations, we demonstrate the superiority of our proposed framework over baseline approaches in question answering tasks. Moreover, our framework exhibits robust performance in handling temporal-based question answering tasks, showcasing its ability to effectively manage time-dependent information.