Exploring Model Kinship for Merging Large Language Models

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Hu, Yedi, Yao, Yunzhi, Zhang, Ningyu, Chen, Huajun, Deng, Shumin
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918145996881920
author Hu, Yedi
Yao, Yunzhi
Zhang, Ningyu
Chen, Huajun
Deng, Shumin
author_facet Hu, Yedi
Yao, Yunzhi
Zhang, Ningyu
Chen, Huajun
Deng, Shumin
contents Model merging has emerged as a key technique for enhancing the capabilities and efficiency of Large Language Models (LLMs). The open-source community has driven model evolution by iteratively merging existing models, yet a principled understanding of the gains and underlying factors in model merging remains limited. In this work, we study model evolution through iterative merging, drawing an analogy to biological evolution, and introduce the concept of model kinship, the degree of similarity or relatedness between LLMs. Through comprehensive empirical analysis, we show that model kinship is closely linked to the performance improvements achieved by merging, providing a useful criterion for selecting candidate models. Building on this insight, we propose a new model merging strategy: Top-k Greedy Merging with Model Kinship, which can improve benchmark performance. Specifically, we discover that incorporating model kinship as a guiding criterion enables continuous merging while mitigating performance degradation caused by local optima, thereby facilitating more effective model evolution. Code is available at https://github.com/zjunlp/ModelKinship.
format Preprint
id arxiv_https___arxiv_org_abs_2410_12613
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Exploring Model Kinship for Merging Large Language Models
Hu, Yedi
Yao, Yunzhi
Zhang, Ningyu
Chen, Huajun
Deng, Shumin
Computation and Language
Artificial Intelligence
Computer Vision and Pattern Recognition
Machine Learning
Multiagent Systems
Model merging has emerged as a key technique for enhancing the capabilities and efficiency of Large Language Models (LLMs). The open-source community has driven model evolution by iteratively merging existing models, yet a principled understanding of the gains and underlying factors in model merging remains limited. In this work, we study model evolution through iterative merging, drawing an analogy to biological evolution, and introduce the concept of model kinship, the degree of similarity or relatedness between LLMs. Through comprehensive empirical analysis, we show that model kinship is closely linked to the performance improvements achieved by merging, providing a useful criterion for selecting candidate models. Building on this insight, we propose a new model merging strategy: Top-k Greedy Merging with Model Kinship, which can improve benchmark performance. Specifically, we discover that incorporating model kinship as a guiding criterion enables continuous merging while mitigating performance degradation caused by local optima, thereby facilitating more effective model evolution. Code is available at https://github.com/zjunlp/ModelKinship.
title Exploring Model Kinship for Merging Large Language Models
topic Computation and Language
Artificial Intelligence
Computer Vision and Pattern Recognition
Machine Learning
Multiagent Systems
url https://arxiv.org/abs/2410.12613