EvoEdit: Lifelong Free-Text Knowledge Editing through Latent Perturbation Augmentation and Knowledge-driven Parameter Fusion

TL;DR

EvoEdit introduces a lifelong free-text knowledge editing method for large language models, enabling natural language updates and continual learning while mitigating forgetting, supported by a new benchmark and multi-level evaluation framework.

Contribution

The paper proposes LF-Edit, a new lifelong free-text knowledge editing task, and introduces EvoEdit, a novel method with Latent Perturbation Augmentation and Parameter Fusion, along with a large benchmark and evaluation framework.

Findings

EvoEdit outperforms existing methods on LF-Edit tasks.

The benchmark MRLF-Bench contains 16,835 free-text edit requests.

Multi-rank evaluation assesses memorization, understanding, and reasoning.

Abstract

Adjusting the outdated knowledge of large language models (LLMs) after deployment remains a major challenge. This difficulty has spurred the development of knowledge editing, which seeks to accurately and efficiently modify a model's internal (parametric) knowledge without retraining it from scratch. However, existing methods suffer from two limitations. First, they depend on structured triplets that are misaligned with the free-text nature of LLM pretraining and fail to capture the nuanced relationships among facts. Second, they typically support one-time knowledge updates, with relatively limited research on the problem of sequential or lifelong editing. To address these gaps, we propose a new task, Lifelong Free-text Knowledge Editing (LF-Edit), which enables models to incorporate updates expressed in natural language and supports continual editing over time. Despite its promise, LF-Edit faces the dual challenge of integrating new knowledge while mitigating the forgetting of prior information. To foster research on this new task, we construct a large-scale benchmark, Multi-Rank Lifelong Free-text Editing Benchmark (MRLF-Bench), containing 16,835 free-text edit requests. We further design a cognitively inspired multi-rank evaluation framework encompassing four levels: memorization, understanding, constrained comprehension, and reasoning. To tackle the challenges inherent in LF-Edit, we introduce a novel approach named EvoEdit that enhances knowledge injection through Latent Perturbation Augmentation and preserves prior information via Knowledge-driven Parameter Fusion. Experimental results demonstrate that EvoEdit substantially outperforms existing knowledge editing methods on the proposed LF-Edit task.