Retention analysis of edited knowledge after fine-tuning

• URL: http://arxiv.org/abs/2507.14198v2
• Date: Fri, 24 Oct 2025 16:26:00 GMT
• Title: Retention analysis of edited knowledge after fine-tuning
• Authors: Fufang Wen, Shichang Zhang,
• Abstract summary: Large language models (LLMs) store vast amounts of knowledge, which often requires updates to correct factual errors, incorporate newly acquired information, or adapt model behavior.<n>Model editing methods have emerged as efficient solutions for such updates, offering localized and precise knowledge modification at significantly lower computational cost than continual training.<n>However, the effect of fine-tuning on previously edited knowledge remains poorly understood.
• Score: 5.1877231178075425
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Large language models (LLMs) store vast amounts of knowledge, which often requires updates to correct factual errors, incorporate newly acquired information, or adapt model behavior. Model editing methods have emerged as efficient solutions for such updates, offering localized and precise knowledge modification at significantly lower computational cost than continual training. In parallel, LLMs are frequently fine-tuned for a wide range of downstream tasks. However, the effect of fine-tuning on previously edited knowledge remains poorly understood. In this work, we systematically investigate how different fine-tuning objectives interact with various model editing techniques. Our findings show that edited knowledge is substantially more susceptible to forgetting during fine-tuning than intrinsic knowledge acquired through pre-training. This analysis highlights a key limitation of current editing approaches and suggests that evaluating edit robustness under downstream fine-tuning is critical for their practical deployment. We further find that knowledge retention can be significantly improved by either augmenting edit knowledge with paraphrases or by freezing layers associated with edited content in fine-tuning stage, offering insight for developing more robust editing algorithms.

Related papers

• Consistency-Aware Editing for Entity-level Unlearning in Language Models [53.522931419965424]
  We introduce a novel consistency-aware editing (CAE) framework for entity-level unlearning.<n>CAE aggregates a diverse set of prompts related to a target entity, including its attributes, relations, and adversarial paraphrases.<n>It then jointly learns a low-rank update guided by a consistency regularizer that aligns the editing directions across prompts.
  arXiv  Detail & Related papers  (2025-12-19T15:18:07Z)
• An Information-Theoretic Framework for Robust Large Language Model Editing [17.984683741974063]
  Large Language Models (LLMs) have become indispensable tools in science, technology, and society.<n>Errors or outdated information within these models can undermine their accuracy and restrict their safe deployment.<n>We introduce a novel framework for editing LLMs, grounded in information bottleneck theory.<n>We present the Information Bottleneck Knowledge Editor (IBKE), which leverages compact latent representations to guide gradient-based updates.
  arXiv  Detail & Related papers  (2025-12-18T06:21:17Z)
• SAKE: Towards Editing Auditory Attribute Knowledge of Large Audio-Language Models [96.81401797908835]
  We introduce SAKE, the first benchmark specifically designed for editing auditory attribute knowledge in Large Audio-Language Models.<n>We benchmark seven editing methods on two LALMs along four dimensions: reliability, generality, audio/text locality, and portability.<n>Results highlight challenges such as preserving intra-attribute knowledge unrelated to the edit, generalizing edits to multimodal reasoning, and maintaining edits under sequential updates.
  arXiv  Detail & Related papers  (2025-10-19T16:22:09Z)
• STEAM: A Semantic-Level Knowledge Editing Framework for Large Language Models [10.320062273796331]
  Large Language Models store extensive factual knowledge acquired during large-scale pre-training.<n>Knowledge editing has emerged as a promising solution for updating outdated or incorrect facts without full retraining.<n>We propose textscSteam, a semantic-level knowledge editing framework that enhances integration of updated knowledge into the model's knowledge structure.
  arXiv  Detail & Related papers  (2025-10-12T01:25:13Z)
• Aligning Language Models with Real-time Knowledge Editing [11.503574001763246]
  We introduce CRAFT, an ever-evolving real-world benchmark for knowledge editing.<n>It features well-designed paired edits for composite reasoning, and evaluates models on alias portability and temporal and common-sense locality.<n> Towards flexible real-time editing, we propose KEDAS, a novel paradigm of knowledge editing alignment featuring diverse edit augmentation and self-adaptive post-alignment inference.
  arXiv  Detail & Related papers  (2025-08-02T10:25:36Z)
• Model Merging for Knowledge Editing [53.799891745131724]
  Large Language Models (LLMs) require continuous updates to maintain accurate and current knowledge as the world evolves.<n>Existing knowledge editing approaches offer various solutions for knowledge updating, but they often struggle with sequential editing scenarios.<n>This paper proposes a two-stage framework combining robust supervised fine-tuning (R-SFT) with model merging for knowledge editing.
  arXiv  Detail & Related papers  (2025-06-14T07:42:39Z)
• ThinkEval: Practical Evaluation of Knowledge Leakage in LLM Editing using Thought-based Knowledge Graphs [3.9295613363026174]
  We present ThinkEval, a framework to quantify indirect knowledge leakage and ripple effects in model-editing.<n>ThinkEval builds and employs specialized knowledge graphs to analyze the causal structure of facts before and after editing.<n>We evaluate five editing techniques: AlphaEdit, RECT, ROME, MEMIT, and PRUNE.
  arXiv  Detail & Related papers  (2025-06-02T07:24:12Z)
• CaKE: Circuit-aware Editing Enables Generalizable Knowledge Learners [88.35958039968081]
  CaKE (Circuit-aware Knowledge Editing) is a novel method that enables more effective integration of updated knowledge in large language models.<n>Results show that CaKE enables more accurate and consistent use of updated knowledge across related reasoning tasks.
  arXiv  Detail & Related papers  (2025-03-20T17:14:34Z)
• Knowledge Updating? No More Model Editing! Just Selective Contextual Reasoning [38.018263569983226]
  We provide an evaluation of ten model editing methods along four dimensions: reliability, generalization, locality, and portability.<n>We then propose a straightforward method called Selective Contextual Reasoning (SCR) for knowledge updating.
  arXiv  Detail & Related papers  (2025-03-07T08:04:25Z)
• Precise Localization of Memories: A Fine-grained Neuron-level Knowledge Editing Technique for LLMs [47.06544781855325]
  We propose a Fine-grained Neuron-level Knowledge Editing (FiNE) method that enhances editing locality without affecting success rates.<n>By precisely identifying and modifying specific neurons within feed-forward networks, FiNE significantly improves knowledge localization and editing.
  arXiv  Detail & Related papers  (2025-03-03T01:30:28Z)
• The Mirage of Model Editing: Revisiting Evaluation in the Wild [70.17413507444704]
  We introduce QAEdit, a new benchmark aligned with widely used question answering (QA) datasets, and WILD, a task-agnostic evaluation framework.<n>Our single editing experiments show that current editing methods perform substantially worse than previously reported.
  arXiv  Detail & Related papers  (2025-02-16T15:57:55Z)
• K-Edit: Language Model Editing with Contextual Knowledge Awareness [71.73747181407323]
  Knowledge-based model editing enables precise modifications to the weights of large language models.<n>We present K-Edit, an effective approach to generating contextually consistent knowledge edits.
  arXiv  Detail & Related papers  (2025-02-15T01:35:13Z)
• Related Knowledge Perturbation Matters: Rethinking Multiple Pieces of Knowledge Editing in Same-Subject [49.559994791305535]
  Current state-of-the-art editing methods struggle when tasked with editing multiple related knowledge pieces for the same subject.<n>We introduce the $textS2textRKE$(Same-Subject Related Knowledge Editing) benchmark.<n>Our experiments reveal that only mainstream locate-then-edit methods, such as ROME and MEMIT, exhibit "related knowledge perturbation"
  arXiv  Detail & Related papers  (2025-02-08T04:47:17Z)
• Uncovering Overfitting in Large Language Model Editing [35.55260822503773]
  We identify and investigate the phenomenon of Editing Overfit, where edited models assign disproportionately high probabilities to the edit target.<n>We propose Learn the Inference (LTI), which introduce a Multi-stage Inference Constraint module to guide the edited models in recalling new knowledge.
  arXiv  Detail & Related papers  (2024-10-10T11:09:00Z)
• Editing the Mind of Giants: An In-Depth Exploration of Pitfalls of Knowledge Editing in Large Language Models [26.516571783335824]
  Recent studies have identified side effects, such as knowledge distortion and the deterioration of general abilities, that have emerged after editing. This survey presents a comprehensive study of these side effects, providing a unified perspective on the challenges of knowledge editing in large language models.
  arXiv  Detail & Related papers  (2024-06-03T15:28:21Z)
• A Comprehensive Study of Knowledge Editing for Large Language Models [82.65729336401027]
  Large Language Models (LLMs) have shown extraordinary capabilities in understanding and generating text that closely mirrors human communication. This paper defines the knowledge editing problem and provides a comprehensive review of cutting-edge approaches. We introduce a new benchmark, KnowEdit, for a comprehensive empirical evaluation of representative knowledge editing approaches.
  arXiv  Detail & Related papers  (2024-01-02T16:54:58Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.