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.
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| Language Understanding | MMLU (test) | -- | 167 | |
| Mathematical Reasoning | GSM8K (val) | Accuracy49.5 | 108 | |
| Multitask Language Understanding | MMLU (val) | Accuracy64 | 94 | |
| Multi-task Language Understanding | MMLU (test) | Normalized Accuracy57.6 | 87 | |
| Mathematical Reasoning | GSM8K (test) | Accuracy (ACC)24.8 | 62 | |
| Common Sense Reasoning | HELLASWAG (test) | Accuracy59.4 | 56 | |
| Commonsense Reasoning | HellaSwag (val) | Accuracy66 | 54 | |
| Natural Language Understanding | GLUE | SST-290 | 40 | |
| Image Classification | Vision Datasets 20 tasks 1.0 (test) | Average Accuracy97.83 | 35 | |
| Truthfulness Evaluation | TruthfulQA (test) | -- | 30 |