Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

EvoGM: Learning to Merge LLMs via Evolutionary Generative Optimization

About

Evolutionary model merging provides a powerful framework for the automated, training-free composition of LLMs through parameter-space search. However, existing methods predominantly rely on stochastic, hand-crafted operators that overlook the underlying performance landscape of the coefficient space. We propose Evolutionary Generative Merging (EvoGM), a framework that transcends manual heuristics by employing learnable generative modeling to optimize merging coefficients. Specifically, EvoGM features a dual-generator architecture with cycle-consistent learning to adaptively sample and refine promising merging candidates. By constructing winner-loser pairs from historical search trajectories, our framework effectively captures high-performance parameter distributions and maximizes data efficiency. This generative process is seamlessly integrated into a multi-round evolutionary pipeline, where elite merged models iteratively serve as new expert foundations. Extensive experiments across diverse benchmarks demonstrate that EvoGM significantly outperforms state-of-the-art baselines, exhibiting robust performance on both seen and unseen tasks. Code and data are available at https://github.com/JiangTao97/evogm.

Tao Jiang, Xinmeng Yu, Chenhao Yi, Yiling Wu, Yan Li, Ran Cheng, Dongmei Jiang, Jianguo Zhang• 2026

Related benchmarks

TaskDatasetResultRank
Language UnderstandingMMLU (test)--
167
Mathematical ReasoningGSM8K (val)
Accuracy49.5
108
Multitask Language UnderstandingMMLU (val)
Accuracy64
94
Multi-task Language UnderstandingMMLU (test)
Normalized Accuracy57.6
87
Mathematical ReasoningGSM8K (test)
Accuracy (ACC)24.8
62
Common Sense ReasoningHELLASWAG (test)
Accuracy59.4
56
Commonsense ReasoningHellaSwag (val)
Accuracy66
54
Natural Language UnderstandingGLUE
SST-290
40
Image ClassificationVision Datasets 20 tasks 1.0 (test)
Average Accuracy97.83
35
Truthfulness EvaluationTruthfulQA (test)--
30
Showing 10 of 26 rows

Other info

Follow for update