{"ID":2853442,"CreatedAt":"2026-06-01T04:54:23.091178241Z","UpdatedAt":"2026-06-01T04:54:23.091178241Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2510.17890","arxiv_id":"2510.17890","title":"MIN-Merging: Merge the Important Neurons for Model Merging","abstract":"Recent advances in deep learning have led to a surge of open-source models across diverse domains. While model merging offers a promising way to combine their strengths, existing approaches often suffer from parameter conflicts that degrade performance on domain-specific tasks. We propose MIN-Merging, a router-based framework that selectively merges the most important neurons to reduce such conflicts. Extensive experiments on Computer Vision(CV) and Natural Language Processing(NLP) benchmarks show that MIN-Merging achieves consistent gains on in-domain tasks while retaining the generalization ability of pretrained models on out-of-domain tasks. These results highlight its effectiveness as a practical solution to the parameter conflict problem in model merging.","short_abstract":"Recent advances in deep learning have led to a surge of open-source models across diverse domains. While model merging offers a promising way to combine their strengths, existing approaches often suffer from parameter conflicts that degrade performance on domain-specific tasks. We propose MIN-Merging, a router-based fr...","url_abs":"https://arxiv.org/abs/2510.17890","url_pdf":"https://arxiv.org/pdf/2510.17890v2","authors":"[\"Yunfei Liang\"]","published":"2025-10-18T15:23:36Z","proceeding":"cs.LG","tasks":"[\"cs.LG\",\"cs.AI\"]","methods":"[]","has_code":false}