Lifelong Knowledge Editing for LLMs with Retrieval-Augmented Continuous Prompt Learning

TL;DR

RECIPE is a continual prompt learning approach that enhances knowledge editing in large language models by integrating retrieval and dynamic prompts, achieving efficient, reliable, and scalable lifelong editing without significant performance loss.

Contribution

The paper introduces RECIPE, a novel retrieval-augmented continuous prompt learning method that improves lifelong knowledge editing in LLMs by addressing efficiency and forgetting issues.

Findings

RECIPE outperforms existing methods in editing accuracy.

It maintains overall LLM performance during continuous editing.

RECIPE achieves faster editing and inference speeds.

Abstract

Model editing aims to correct outdated or erroneous knowledge in large language models (LLMs) without the need for costly retraining. Lifelong model editing is the most challenging task that caters to the continuous editing requirements of LLMs. Prior works primarily focus on single or batch editing; nevertheless, these methods fall short in lifelong editing scenarios due to catastrophic knowledge forgetting and the degradation of model performance. Although retrieval-based methods alleviate these issues, they are impeded by slow and cumbersome processes of integrating the retrieved knowledge into the model. In this work, we introduce RECIPE, a RetriEval-augmented ContInuous Prompt lEarning method, to boost editing efficacy and inference efficiency in lifelong learning. RECIPE first converts knowledge statements into short and informative continuous prompts, prefixed to the LLM's input query embedding, to efficiently refine the response grounded on the knowledge. It further integrates the Knowledge Sentinel (KS) that acts as an intermediary to calculate a dynamic threshold, determining whether the retrieval repository contains relevant knowledge. Our retriever and prompt encoder are jointly trained to achieve editing properties, i.e., reliability, generality, and locality. In our experiments, RECIPE is assessed extensively across multiple LLMs and editing datasets, where it achieves superior editing performance. RECIPE also demonstrates its capability to maintain the overall performance of LLMs alongside showcasing fast editing and inference speed.