mlsanityData to 13 Oct 2026

WilKE: Wise-Layer Knowledge Editor for Lifelong Knowledge Editing

arXiv 2402.10987 ↗First posted 16 Feb 2024cs.CL cs.AIcited by 40

Not on the map. Why?

Abstract (arXiv)

Knowledge editing aims to rectify inaccuracies in large language models (LLMs) without costly retraining for outdated or erroneous knowledge. However, current knowledge editing methods primarily focus on single editing, failing to meet the requirements for lifelong editing. This study reveals a performance degradation encountered by knowledge editing in lifelong editing, characterized by toxicity buildup and toxicity flash, with the primary cause identified as pattern unmatch. We introduce a knowledge editing approach named Wise-Layer Knowledge Editor (WilKE), which selects editing layer based on the pattern matching degree of editing knowledge across different layers in language models. Experimental results demonstrate that, in lifelong editing, WilKE exhibits an average improvement of 46.2% and 67.8% on editing GPT2-XL and GPT-J relative to state-of-the-art knowledge editing methods.

Key facts. Cited by 40 papers in the map (data through 13 Oct 2026). Builds on 20 of its 29 references that are in the map. First posted on arXiv 16 Feb 2024.

Builds on

20 of 29 references.

References · 20

  1. 2020Language Models are Few-Shot Learnerscited by 32,320
  2. 2023GPT-4 Technical Reportcited by 18,278
  3. 2022Locating and Editing Factual Associations in GPTcited by 2,650
  4. 2020Transformer Feed-Forward Layers Are Key-Value Memoriescited by 1,323
  5. 2022Mass-Editing Memory in a Transformercited by 990
  6. 2022Toy Models of Superpositioncited by 950
  7. 2021Knowledge Neurons in Pretrained Transformerscited by 597
  8. 2021Editing Factual Knowledge in Language Modelscited by 551
  9. 2021Fast Model Editing at Scalecited by 493
  10. 2022Memory-Based Model Editing at Scalecited by 446
  11. 2023Editing Large Language Models: Problems, Methods, and Opportunitiescited by 370
  12. 2022Aging with GRACE: Lifelong Model Editing with Discrete Key-Value Adaptorscited by 270
8 more
  1. 2023Transformer-Patcher: One Mistake worth One Neuroncited by 197
  2. 2023PMET: Precise Model Editing in a Transformercited by 172
  3. 2023EasyEdit: An Easy-to-use Knowledge Editing Framework for Large Language Modelscited by 97
  4. 2019Augmenting Self-attention with Persistent Memorycited by 93
  5. 2022Calibrating Factual Knowledge in Pretrained Language Modelscited by 92
  6. 2023Massive Editing for Large Language Models via Meta Learningcited by 62
  7. 2023Unveiling the Pitfalls of Knowledge Editing for Large Language Modelscited by 43
  8. 2023Evaluating the Reversal Curse in Model Editingcited by 34

Built on by

Cited by year (2026 to 13 Oct)

Citing papers per year: 2024: 20; 2025: 14; 2026: 6.

40 papers.

Sort:

By area:

  1. 2024AlphaEdit: Null-Space Constrained Knowledge Editing for Language Modelscited by 232
  2. 2024Continual Learning of Large Language Models: A Comprehensive Surveycited by 230
  3. 2025Lifelong Learning of Large Language Model based Agents: A Roadmapcited by 81
  4. 2024Rebuilding ROME : Resolving Model Collapse during Sequential Model Editingcited by 50
  5. 2024Lifelong Knowledge Editing for LLMs with Retrieval-Augmented Continuous Prompt Learningcited by 40
  6. 2024AKEW: Assessing Knowledge Editing in the Wildcited by 36
  7. 2024Multi-hop Question Answering under Temporal Knowledge Editingcited by 36
  8. 2024Locate-then-edit for Multi-hop Factual Recall under Knowledge Editingcited by 36
  9. 2024Perturbation-Restrained Sequential Model Editingcited by 27
  10. 2024Attribution Analysis Meets Model Editing: Advancing Knowledge Correction in Vision Language Models with VisEditcited by 14
  11. 2024Lifelong Knowledge Editing for Vision Language Models with Low-Rank Mixture-of-Expertscited by 14
  12. 2024Leveraging Logical Rules in Knowledge Editing: A Cherry on the Topcited by 12
28 more
  1. 2024OneEdit: A Neural-Symbolic Collaboratively Knowledge Editing Systemcited by 12
  2. 2025CoME: An Unlearning-based Approach to Conflict-free Model Editingcited by 12
  3. 2025SAKE: Steering Activations for Knowledge Editingcited by 12
  4. 2025Precise In-Parameter Concept Erasure in Large Language Modelscited by 11
  5. 2024Consecutive Batch Model Editing with HooK Layerscited by 9
  6. 2024Knowledge in Superposition: Unveiling the Failures of Lifelong Knowledge Editing for Large Language Modelscited by 8
  7. 2024ELDER: Enhancing Lifelong Model Editing with Mixture-of-LoRAcited by 8
  8. 2024O-Edit: Orthogonal Subspace Editing for Language Model Sequential Editingcited by 8
  9. 2025UltraEdit: Training-, Subject-, and Memory-Free Lifelong Editing in Language Modelscited by 8
  10. 2024Enhancing Data Privacy in Large Language Models through Private Association Editingcited by 6
  11. 2025Is Fine-Tuning an Effective Solution? Reassessing Knowledge Editing for Unstructured Datacited by 6
  12. 2025Revealing the Deceptiveness of Knowledge Editing: A Mechanistic Analysis of Superficial Editingcited by 5
  13. 2024ComprehendEdit: A Comprehensive Dataset and Evaluation Framework for Multimodal Knowledge Editingcited by 4
  14. 2024Cracking Factual Knowledge: A Comprehensive Analysis of Degenerate Knowledge Neurons in Large Language Modelscited by 3
  15. 2025SetKE: Knowledge Editing for Knowledge Elements Overlapcited by 3
  16. 2024Keys to Robust Edits: from Theoretical Insights to Practical Advancescited by 2
  17. 2025Constraining Sequential Model Editing with Editing Anchor Compressioncited by 2
  18. 2025On the Superimposed Noise Accumulation Problem in Sequential Knowledge Editing of Large Language Modelscited by 2
  19. 2025EvoEdit: Lifelong Free-Text Knowledge Editing through Latent Perturbation Augmentation and Knowledge-driven Parameter Fusioncited by 2
  20. 2026Multiplicative Orthogonal Sequential Editing for Language Modelscited by 2
  21. 2026Mitigating Memorization Where It Happenscited by 2
  22. 2026Long-Tail Knowledge in Large Language Models: Taxonomy, Mechanisms, Interventions and Implicationscited by 2
  23. 2025QueueEDIT: Structural Self-Correction for Sequential Model Editing in LLMscited by 1
  24. 2026Reforming the Mechanism: Editing Reasoning Patterns in LLMs with Circuit Reshapingcited by 1
  25. 2026Towards Scalable Lifelong Knowledge Editing with Selective Knowledge Suppressioncited by 1
  26. 2025Towards Atoms of Large Language Modelscited by 0
  27. 2025An Information-Theoretic Framework for Robust Large Language Model Editingcited by 0
  28. 2026X-Edit: Exact, Explicit, and Explainable Null-Space Editing for Medical Vision Transformerscited by 0