Knowledge Editing for Large Language Model with Knowledge Neuronal Ensemble

TL;DR

This paper introduces Knowledge Neuronal Ensemble (KNE), a novel method for editing knowledge in large language models that improves precision, accuracy, and dynamic interaction among neurons, outperforming existing methods on key metrics.

Contribution

The paper proposes KNE, a new knowledge editing approach that groups neurons into ensembles to enhance localization and interaction, addressing limitations of prior methods.

Findings

KNE significantly improves knowledge editing accuracy.

KNE outperforms baseline methods on portability and locality metrics.

Experimental results validate the effectiveness of the proposed approach.

Abstract

As real-world knowledge is constantly evolving, ensuring the timeliness and accuracy of a model's knowledge is crucial. This has made knowledge editing in large language models increasingly important. However, existing knowledge editing methods face several challenges, including parameter localization coupling, imprecise localization, and a lack of dynamic interaction across layers. In this paper, we propose a novel knowledge editing method called Knowledge Neuronal Ensemble (KNE). A knowledge neuronal ensemble represents a group of neurons encoding specific knowledge, thus mitigating the issue of frequent parameter modification caused by coupling in parameter localization. The KNE method enhances the precision and accuracy of parameter localization by computing gradient attribution scores for each parameter at each layer. During the editing process, only the gradients and losses associated with the knowledge neuronal ensemble are computed, with error backpropagation performed accordingly, ensuring dynamic interaction and collaborative updates among parameters. Experimental results on three widely used knowledge editing datasets show that the KNE method significantly improves the accuracy of knowledge editing and achieves, or even exceeds, the performance of the best baseline methods in portability and locality metrics.