Knowledge Editing

Ask about this topic

Overview

Knowledge editing aims to correct or add information in a model without retraining it from scratch, so models can adapt to new facts, rules, and user-specific information. Approaches include targeted weight updates, low-rank adapters and null-space constraints, as well as external memories and retrieval systems that supply updates at inference time. Researchers test whether edits work across paraphrases, related questions, and multi-hop reasoning while measuring locality, capability retention, and stability over long sequences of updates. A central tension is that strong performance on direct edit queries may reflect surface compliance rather than changed internal knowledge, and edits can interfere with other facts, reasoning, or safety behavior over time.

Latest papers

CardsList

86 papers11 papers in the last four weeks, level with the four weeks before. 0.1% of all new papers.

  1. KARMA: Training-Free Knowledge-Atomic Harness Evolution for Industrial Advertising Agents

    Oct 29, 2026Xinhua Miao, Yuanzhen Rong, Weiru Zhang +1LLM AgentsSelf-Improving Agents

    Alibaba International Digital Commerce Group, Hangzhou, China

  2. Evaluating Cultural Knowledge Editing in Vision-Language Models

    Oct 29, 2026Malvina Nikandrou, Georgios Pantazopoulos, Ioannis Konstas +1Vision-Language ModelsKnowledge Editing

    University of Edinburgh · Heriot-Watt University

  3. Weakness Localization Is Not Enough: State-Aware Recursive Knowledge Improvement for Large Language Models

    Oct 27, 2026Runze Jiang, Yan Wang, Boran Jin +2Knowledge Augmentation for Language ModelsKnowledge Editing

    Peking University · Alibaba Group · Delft University of Technology

  4. DynEdit: Dynamic Entropy-Guided Sequential Editing for Large Language Models

    Oct 26, 2026Jinhu Fu, Yan Bai, Yihang Lou +4Knowledge EditingLanguage Model Editing

    Beijing University of Posts and Telecommunications · Peking University · AMAP, Alibaba Group +1

  5. TaxoSAE: Rethinking Sparse Autoencoders through Taxonomic Logic

    Oct 22, 2026Xinyu Guo, Minglai Yang, Clayton T. Morrison +2Sparse AutoencodersHierarchical Representation Learning

    University of Arizona · Scale AI

  6. Diagnosing and Mitigating Evidence Misgrounding in Multimodal Knowledge Editing

    Oct 21, 2026Zhiwu Wang, Yichen Wu, Jinqian Yang +4Multimodal Knowledge EditingKnowledge Editing

    School of Mathematics and Statistics, Xi’an Jiaotong University, Xi’an, Shaanxi, China · Harvard Medical School, Boston, MA 02155, U.S. · School of Mathematics and Statistics, Xidian University, Xi’an, Shaanxi, China +2

  7. FisherEdit: Fisher Information Guided Knowledge Editing for Large Language Models

    Oct 19, 2026Yizhou Yang, Jiahong Fu, Zhiwu Wang +6Knowledge EditingLanguage Model Editing

    Xi’an Jiaotong University · Zhejiang University · City University of Hong Kong +2

  8. Mixture-of-Memories: Low-Overhead Language Modeling with Modular Knowledge Injection

    Oct 18, 2026Yifan Lu, Rubin Wei, Ziyun Zou +8Memory-Augmented Language ModelsKnowledge Augmentation for Language Models

    MBZUAI, Abu Dhabi, UAE · LUMIA Lab, School of Artificial Intelligence, Shanghai Jiao Tong University, China · The Chinese University of Hong Kong, Hong Kong +1

  9. Writability and the Pathway: A Controlled Study of Knowledge Injection into Addressed Memory

    Oct 15, 2026Wei Gao, Cheng Tian, Duan Yang +1Knowledge Augmentation for Language ModelsKnowledge Editing

    Shibei Qiyuan AI Technology · Xi'an Jiaotong University

  10. Cortex: Consolidating Context into Parametric Memory

    Oct 14, 2026Jintian Zhang, Feng He, Xinnian Liang +5Continual Learning for LLMsKnowledge Editing

    Zhejiang University · ByteDance Seed · Tsinghua University

  11. Do correct answers have stable latent knowledge support?

    Oct 11, 2026Chuancheng Shi, Shangze Li, Wenhua Wu +4LLM EvaluationRepresentation Geometry in Language Models

    University of Sydney Sydney, Australia · Nanjing University of Science and Technology Nanjing, China · Edith Cowan University Perth, Australia +1

  12. Not Every Change Is Necessary: Recoverable Drift in Large Language Model Unlearning

    Oct 8, 2026Xunlei Chen, Qinghui Gong, Jingkun Xue +3Machine UnlearningLLM Unlearning

    University of Electronic Science and Technology of China · Southwest Jiaotong University

  13. FedAlphaEdit: Null-Space-Aligned Merging for Collaborative Knowledge Editing

    Oct 8, 2026Sota Sugawara, Yukihiko OkadaKnowledge EditingLanguage Model Editing

    Graduate School of Science and Technology, University of Tsukuba, Tsukuba, Japan · Center for Artificial Intelligence Research, Tsukuba Institute for Advanced Research, University of Tsukuba, Tsukuba, Japan · Institute of Systems and Information Engineering, University of Tsukuba, Tsukuba, Japan

  14. EngramEdit: Decoupled Knowledge Updates in LLMs through Conditional Memory

    Oct 7, 2026Hongru Cai, Ran Wei, Wenjie Wang +4Knowledge EditingFactual Knowledge in Language Models

    The Hong Kong Polytechnic University · Hangzhou Diagens Biotechnology Co., Ltd · University of Science and Technology of China

  15. Improving Atomic-Fact Recall via Focused Views in Unstructured Knowledge Editing

    Oct 2, 2026Ding Wu, Ye Zhang, Haoyu Wang +1Knowledge EditingFactual Knowledge in Language Models

    Georgia Institute of Technology · University of Pennsylvania · University at Albany +1

  16. Positions Are Not Facts: The Mismatch Between KV Caches and Memory

    Sep 27, 2026Changhai Zhou, Yuhua Zhou, Shiyang Zhang +5Knowledge EditingKnowledge Conflicts in Language Models

    Fudan University · Baidu · Zhejiang University +2

  17. ManiEdit: Sequential Unstructured Knowledge Editing for Language Models from a Manifold Perspective

    Sep 27, 2026Rui Liu, Chenheng Zhang, Haoxuan Li +1Knowledge EditingContinual Learning for LLMs

    State Key Lab of General Artificial Intelligence, School of Intelligence Science and Technology, Peking University · Institute for Artificial Intelligence, Peking University

  18. ALLOT: Budgeted Hybrid-Memory Routing for Knowledge Updates in LLMs

    Sep 26, 2026Shanfeng Huang, Zhou Fang, Song Xiao +1Memory-Augmented Language ModelsLLM Routing

    Baidu Inc.

  19. Style-Debiased DPO: Updating LLM Knowledge with Factuality-Aware Synthetic Preference Data

    Sep 15, 2026Takayuki Yamamoto, Daisuke KawaharaKnowledge EditingDirect Preference Optimization

    Waseda University

  20. Specification Oracles

    Sep 11, 2026Atticus Cull, Justin McCarthyMemory-Augmented Language ModelsFactual Knowledge in Language Models

    Diffusion Redwood City, California, USA

  21. When Successful Knowledge Graph Edits Displace Correct Answers: Rank-Level Locality beyond Parameter Support

    Sep 10, 2026Yi-Cheng Lai, Jerry Wang, Hsin-Ling Hsu +3Knowledge EditingKG Embedding

    Institute of Information Science, Academia Sinica Taipei, Taiwan · University of Illinois Urbana-Champaign Urbana, IL, USA · National Chengchi University Taipei, Taiwan

  22. Through the Looking Glass: Directly Reading and Writing Transformers

    Sep 9, 2026Mark OskinKnowledge EditingTransformer Interpretability

    School of Computer Science and Engineering University of Washington

  23. Popular Knowledge Propagates More Errors in LLM Knowledge Updating

    Sep 8, 2026Yuji Zhang, Weibing Wang, Cheng Qian +5Knowledge EditingFactual Knowledge in Language Models

    University of Illinois Urbana-Champaign · City University of New York · Massachusetts Institute of Technology +1

  24. Selective Knowledge Edit Reversal via Gated Singular Vector Shrinkage

    Sep 2, 2026Weifeng Jiang, Ruirui Chen, Qianren Mao +3Knowledge EditingLLM Safety

    College of Computing and Data Science, Nanyang Technological University, Singapore. · Digital Trust Centre, Nanyang Technological University, Singapore. · Institute of Advanced Intelligence and Computing, Agency for Science, Technology and Research, Singapore. +2

  25. Parallel Universes for Temporal Evaluation and Knowledge Updating in LLMs

    Aug 31, 2026Jonathan Zheng, Zirui Shao, Alan Ritter +1Synthetic Data PretrainingKnowledge Editing

    Georgia Institute of Technology · Zhejiang University

  26. Sequential knowledge editing breaks a model's ability to tell good evidence from bad, without costing it accuracy

    Aug 31, 2026Atul AnandKnowledge EditingLanguage Model Editing

  27. KLOD: Locality-Preserving Knowledge Editing via Non-Target Distribution Preservation

    Aug 28, 2026Hojun Jeong, Gyunyeop Kim, Sangwoo KangKnowledge EditingLanguage Model Editing

    Department of Computing, Gachon University

  28. On Scope Classification and Current Knowledge-Editing Benchmarks: A Negative Result, with INLAY as a Gradient-Free Case Study

    Aug 26, 2026Aditya Pratap SinghKnowledge Editing

    Independent Researcher

  29. Towards Reliable, Generalizable, and Specific In-Context Knowledge Editing via Multi-Objective Reinforcement Learning

    Aug 25, 2026Xuzhong Wang, Maiqi Jiang, Tejal Nair +3Knowledge EditingLanguage Model Editing

    College of William and Mary · Tribhuvan University

  30. In-Cell Learning: Language Models That Update Their Own Weights in Sequence Without Changing the File They Ship

    Aug 21, 2026Zifeng Liu, Yaxin Lu, Xuanhan Wu +6LLM QuantizationKnowledge Editing

    Big Data and Artificial Intelligence Center, The Third Affiliated Hospital of Sun Yat-sen University · Institute for Frontier Interdisciplinary Research in Health Sciences and Technology, Sun Yat-sen University · Sun Yat-sen University Institute of Artificial Intelligence +7