A new study from Meta and NYU reveals a two‑step technique—SAMD to locate concept‑specific attention heads and SAMI to scale their influence—enabling precise, low‑cost editing of transformer models for tasks ranging from factual recall to safety control.
This article explains model editing—injecting or updating knowledge in AI models—distinguishes it from post‑training, outlines reliability, generalization and locality metrics, and surveys both parameter‑free (e.g., IKE) and parameter‑based methods such as ROME, hypernetworks, and MEND, highlighting practical challenges.
This article introduces DAFNet, a dynamic auxiliary fusion framework that enables efficient sequential editing of large language models by injecting knowledge with reduced resource costs while preserving model reliability, generalization, and mitigating hallucination, and details its dataset, architecture, and evaluation results.
