FUNDAMENTAL METHODS FOR MODEL EDITING

• Transformer Feed-Forward Layers Are Key-Value Memories - Paper Link, Video Explanation
• Editing Factual Knowledge in Language Models - Paper Link
• Knowledge Neurons in Pretrained Transformers - Paper Link, Video Explanation
• MEND - Fast Model Editing at Scale - Paper Link, Video Explanation
• ROME - Locating and Editing Factual Associations in GPT - Paper Link, Video by original authors
• SERAC - Memory-Based Model Editing at Scale - Paper Link, Video Explanation
• MEMIT - Mass-Editing Memory in a Transformer - Paper Link, Video Explanation
• EMMET - A Unified Framework for Model Editing - Paper Link
• MALMEN - Massive Editing for Large Language Model via Meta Learning - Paper Link

SELECTED ANALYSIS PAPERS FOR MODEL EDITING

• Does Localization Inform Editing? - Paper Link
• Evaluating the Ripple Effects of Knowledge Editing in Language Models - Paper Link
• Unveiling the Pitfalls of Knowledge Editing for Large Language Models - Paper Link
• Model Editing at Scale leads to Gradual and Catastrophic Forgetting - Paper link, Video by original authors
• Editing Large Language Models: Problems, Methods, and Opportunities - Paper Link
• Is Bigger Edit Batch Size Always Better? - An Empirical Study on Model Editing with Llama-3 - Paper Link