paper-with-me

홈 › Papers

Revealing and Mitigating Over-Attention in Knowledge Editing

2025-02-20 · Pinzheng Wang, Zecheng Tang, Keyan Zhou, Juntao Li, Qiaoming Zhu, Min Zhang

Large Language Models have demonstrated superior performance across a wide range of tasks, but they still exhibit undesirable errors due to incorrect knowledge learned from the training data. To avoid this, knowledge editing methods emerged to precisely edit the specific model knowledge via efficiently modifying a very small percentage of parameters. % However, those methods can lead to the problem of Specificity Failure: when the content related to the edited knowledge occurs in the context, it can inadvertently corrupt other pre-existing knowledge. However, those methods can lead to the problem of Specificity Failure, where the existing knowledge and capabilities are severely degraded due to editing. Our preliminary indicates that Specificity Failure primarily stems from the model's attention heads assigning excessive attention scores to entities related to the edited knowledge, thereby unduly focusing on specific snippets within the context, which we denote as the Attention Drift phenomenon. To mitigate such Attention Drift issue, we introduce a simple yet effective method Selective Attention Drift Restriction}(SADR), which introduces an additional regularization term during the knowledge editing process to restrict changes in the attention weight distribution, thereby preventing undue focus on the edited entity. Experiments on five frequently used strong LLMs demonstrate the effectiveness of our method, where SADR can significantly mitigate Specificity Failure in the predominant knowledge editing tasks.

📄 PDF Abstract BibTeX arXiv:2502.14838

Code (1)

PinzhengWang322/Reveal_Attention_Drift 공식 구현 pytorch

Tasks

knowledge editingSpecificity

Methods 이 논문이 사용한 방법론

Softmax The Softmax output function transforms a previous layer's output into a vector of probabilities. It is commonly used for multiclass classification. Given an input vector $x$…
Attention 설명 없음
Focus 설명 없음

Similar Papers 제목 키워드 기반

Commonsense Knowledge Editing Based on Free-Text in LLMs

2024-10-31 · Xiusheng Huang, Yequan Wang, Jun Zhao, Kang Liu

Knowledge editing technology is crucial for maintaining the accuracy and timeliness of large language models (LLMs) . However, the setting of this task overlooks a significant portion of commonsense knowledge based on fr…

knowledge editing

Revealing the Deceptiveness of Knowledge Editing: A Mechanistic Analysis of Superficial Editing

2025-05-19 · Jiakuan Xie, Pengfei Cao, Yubo Chen, Kang Liu 외

Knowledge editing, which aims to update the knowledge encoded in language models, can be deceptive. Despite the fact that many existing knowledge editing algorithms achieve near-perfect performance on conventional metric…

knowledge editing

Towards a Principled Evaluation of Knowledge Editors

2025-07-08 · Sebastian Pohl, Max Ploner, Alan Akbik arxiv

Model editing has been gaining increasing attention over the past few years. For Knowledge Editing in particular, more challenging evaluation datasets have recently been released. These datasets use different methodologi…

knowledge editing

HIME: Mitigating Object Hallucinations in LVLMs via Hallucination Insensitivity Model Editing

2026-02-21 · Ahmed Akl, Abdelwahed Khamis, Ali Cheraghian, Zhe Wang 외 arxiv

Large Vision-Language Models (LVLMs) have demonstrated impressive multimodal understanding capabilities, yet they remain prone to object hallucination, where models describe non-existent objects or attribute incorrect fa…

MultiMedEdit: A Scenario-Aware Benchmark for Evaluating Knowledge Editing in Medical VQA

2025-08-09 · Shengtao Wen, Haodong Chen, Yadong Wang, Zhongying Pan 외 arxiv

Knowledge editing (KE) provides a scalable approach for updating factual knowledge in large language models without full retraining. While previous studies have demonstrated effectiveness in general domains and medical Q…

knowledge editingVisual Reasoning