On the Superimposed Noise Accumulation Problem in Sequential Knowledge Editing of Large Language Models

TL;DR

This paper identifies the superimposed noise accumulation problem in sequential knowledge editing of large language models, and proposes DeltaEdit to mitigate it, improving editing success rates.

Contribution

The paper introduces DeltaEdit, a novel method with dynamic orthogonal constraints to reduce knowledge conflicts and superimposed noise in sequential model editing.

Findings

DeltaEdit reduces superimposed noise significantly.

Achieves 16.8% improvement over the strongest baseline.

Theoretical analysis links noise accumulation to knowledge conflicts.

Abstract

Sequential knowledge editing techniques aim to continuously update knowledge in large language models at low cost, preventing models from generating outdated or incorrect information. However, existing sequential editing methods suffer from a significant decline in editing success rates after long-term editing. Through theoretical analysis and experiments, our findings reveal that as the number of edits increases, the model's output increasingly deviates from the desired target, leading to a drop in editing success rates. We refer to this issue as the superimposed noise accumulation problem. Our further analysis demonstrates that the problem is related to the erroneous activation of irrelevant knowledge and conflicts between activated knowledge. Based on this analysis, a method named DeltaEdit is proposed that reduces conflicts between knowledge through dynamic orthogonal constraint strategies. Experiments show that DeltaEdit significantly reduces superimposed noise, achieving a 16.8% improvement in editing performance over the strongest baseline.