Elamur Research⌕
☰
home / papers / 2026 / Nature-Inspired Population-Based Evolution of Large Language Models

Nature-Inspired Population-Based Evolution of Large Language Models

Paper · 2026 · Yiqun Zhang, Peng Ye, Xiaocui Yang et al.

Summary

The paper frames a new problem: how to improve a whole group of large language models by treating it as an evolving population rather than tuning one model. The authors propose GENOME, which adapts the classic genetic algorithm to LLMs. Model weights, mainly LoRA adapters, play the role of genes, and task accuracy on a small validation set serves as fitness. Crossover blends parent weights, mutation adds sparse random noise, and selection keeps strong individuals. The extended variant, GENOME+, adds two more steps. Succession moves weights toward the best performers and away from the worst, and ensemble combines answers from the top-k models by majority vote. The method needs no gradients and only about 200 labelled examples per new task. Across 12 datasets it beats eight merging and adaptation baselines, including Model Swarms, with the largest gains on reasoning-heavy tasks. It also handles multi-task adaptation, transfers zero-shot to unseen tasks and languages, and runs on one consumer GPU. This makes reuse of existing expert adapters cheap and widely accessible.

Key points

  • GENOME+ beats the strongest initial model by up to 54.8% (relative). It averages +10.75% over Model Swarms, with +36.3% on DROP and +19.3% on MGSM.
  • When adapting to two tasks at once, it beats Model Swarms by 13.46% on average, including +50.9% on DROP. In zero-shot transfer it gains +11.72% over Expert Fusion. On unseen Flores-101 languages, GENOME and GENOME+ gain +12.83% and +15.98% over the best single model.
  • Performance keeps improving as the population grows to 40 models. Results also hold with larger base models, Llama-3.1-8B and Qwen-2.5-14B.
  • In ablations, dropping ensemble costs 11.47% on average. Removing crossover and mutation together costs 12.47%, more than removing either alone.
  • Experiments with gemma-2-2b-it run on a single RTX 4090 (24GB). The method requires an existing pool of LoRA adapters that share the same base architecture, so it cannot merge heterogeneous models. Code is released as GENOME on GitHub.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

 Institutions

🔬Shanghai Artificial Intelligence LaboratoryShanghai · China117 works · 11,281 citations🎓Northeastern UniversityShenyang · China42 works · 1,940 citations

 Fields, topics and concepts

📚 Shelves 4The field of science this work belongs to (OpenAlex fields).Computer Science · 16,667 worksArtificial Intelligence · 16,194 worksArts and Humanities · 16,039 worksLanguage and Linguistics · 15,891 works

🧭 Topics 1Research subjects and languages the work is about.Language models & pretraining · 3,320 works

Cite

Cite as Yiqun Zhang, Peng Ye, Xiaocui Yang, Shi Feng, Shufei Zhang, Lei Bai, Wanli Ouyang and Shuyue Hu. 2026. Nature-Inspired Population-Based Evolution of Large Language Models. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers).

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2026.acl-long.2044

@inproceedings{zhang2026nature,
  title = {{Nature-Inspired Population-Based Evolution of Large Language Models}},
  author = {Yiqun Zhang, Peng Ye, Xiaocui Yang, Shi Feng, Shufei Zhang, Lei Bai, Wanli Ouyang and Shuyue Hu},
  booktitle = {Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)},
  year = {2026},
  doi = {10.18653/v1/2026.acl-long.2044},
  url = {https://doi.org/10.18653/v1/2026.acl-long.2044}
}

References

47 references, 25 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Takuya Akiba, Makoto Shing, Yujin Tang, Qi Sun, and David Ha. 2024. Evolutionary optimization of model merging recipes. arXiv preprint arXiv:2403.13187. — [arXiv:2403.13187](https://arxiv.org/abs/2403.13187) find
  2. Jacob Austin, Augustus Odena, Maxwell Nye, Maarten Bosma, Henryk Michalewski, David Dohan, Ellen Jiang, Carrie Cai, Michael Terry, Quoc Le, and 1 others. 2021. Program synthesis with large language models. arXiv preprint arXiv:2108.07732.
  3. Sahil Chaudhary. 2023. Code alpaca: An instruction-following llama model for code generation. https://github.com/sahil280114/codealpaca. ↗
  4. Shuhao Chen, Weisen Jiang, Baijiong Lin, James T. Kwok, and Yu Zhang. 2024. Routerdc: Query-based router by dual contrastive learning for assembling large language models. Preprint, arXiv:2409.19886. — [arXiv:2409.19886](https://arxiv.org/abs/2409.19886) find
  5. Hyung Won Chung, Le Hou, Shayne Longpre, Barret Zoph, Yi Tay, William Fedus, Yunxuan Li, Xuezhi Wang, Mostafa Dehghani, Siddhartha Brahma, and 1 others. 2024. Scaling instruction-finetuned language models. Journal of Machine Learning Research, 25(70):1–53. find
  6. Peter Clark, Isaac Cowhey, Oren Etzioni, Tushar Khot, Ashish Sabharwal, Carissa Schoenick, and Oyvind Tafjord. 2018. Think you have solved question answering? try arc, the ai2 reasoning challenge. Preprint, arXiv:1803.05457.
  7. Karl Cobbe, Vineet Kosaraju, Mohammad Bavarian, Mark Chen, Heewoo Jun, Lukasz Kaiser, Matthias Plappert, Jerry Tworek, Jacob Hilton, Reiichiro Nakano, Christopher Hesse, and John Schulman. 2021. Training verifiers to solve math word problems. arXiv preprint arXiv:2110.14168.
  8. Dheeru Dua, Yizhong Wang, Pradeep Dasigi, Gabriel Stanovsky, Sameer Singh, and Matt Gardner. 2019. Drop: A reading comprehension benchmark requiring discrete reasoning over paragraphs. Preprint, arXiv:1903.00161.
  9. Shangbin Feng, Zifeng Wang, Yike Wang, Sayna Ebrahimi, Hamid Palangi, Lesly Miculicich, Achin Kulshrestha, Nathalie Rauschmayr, Yejin Choi, Yulia Tsvetkov, and 1 others. 2024. Model swarms: Collaborative search to adapt llm experts via swarm intelligence. arXiv preprint arXiv:2410.11163. — [arXiv:2410.11163](https://arxiv.org/abs/2410.11163) find
  10. Hamish Ivison, Yizhong Wang, Valentina Pyatkin, Nathan Lambert, Matthew Peters, Pradeep Dasigi, Joel Jang, David Wadden, Noah A Smith, Iz Beltagy, and 1 others. 2023b. Camels in a changing climate: Enhancing lm adaptation with tulu 2. arXiv preprint arXiv:2311.10702.
  11. Charles Goddard, Shamane Siriwardhana, Malikeh Ehghaghi, Luke Meyers, Vlad Karpukhin, Brian Benedict, Mark McQuade, and Jacob Solawetz. 2024. Arcee’s mergekit: A toolkit for merging large language models. arXiv preprint arXiv:2403.13257.
  12. Dong-Hwan Jang, Sangdoo Yun, and Dongyoon Han. 2024. Model stock: All we need is just a few fine-tuned models. In European Conference on Computer Vision, pages 207–223. Springer. find
  13. Naman Goyal, Cynthia Gao, Vishrav Chaudhary, Peng-Jen Chen, Guillaume Wenzek, Da Ju, Sanjana Krishnan, Marc’Aurelio Ranzato, Francisco Guzmán, and Angela Fan. 2022. The flores-101 evaluation benchmark for low-resource and multilingual machine translation. Transactions of the Association for Computat… find
  14. Aaron Grattafiori, Abhimanyu Dubey, Abhinav Jauhri, Abhinav Pandey, Abhishek Kadian, Ahmad Al-Dahle, Aiesha Letman, Akhil Mathur, Alan Schelten, Alex Vaughan, and 1 others. 2024. The llama 3 herd of models. arXiv preprint arXiv:2407.21783.
  15. Dan Hendrycks, Collin Burns, Steven Basart, Andy Zou, Mantas Mazeika, Dawn Song, and Jacob Steinhardt. 2021a. Measuring massive multitask language understanding. Proceedings of the International Conference on Learning Representations (ICLR).
  16. Dan Hendrycks, Collin Burns, Saurav Kadavath, Akul Arora, Steven Basart, Eric Tang, Dawn Song, and Jacob Steinhardt. 2021b. Measuring mathematical problem solving with the math dataset. NeurIPS.
  17. John H Holland. 1992. Genetic algorithms. Scientific american, 267(1):66–73. find
  18. Edward J Hu, Yelong Shen, Phillip Wallis, Zeyuan Allen-Zhu, Yuanzhi Li, Shean Wang, Lu Wang, and Weizhu Chen. 2021. Lora: Low-rank adaptation of large language models. arXiv preprint arXiv:2106.09685.
  19. Yedi Hu, Yunzhi Yao, Ningyu Zhang, Shumin Deng, and Huajun Chen. 2024. Exploring model kinship for merging large language models. arXiv preprint arXiv:2410.12613. find
  20. Andreas Köpf, Yannic Kilcher, Dimitri von Rütte, Sotiris Anagnostidis, Zhi Rui Tam, Keith Stevens, Abdullah Barhoum, Duc Nguyen, Oliver Stanley, Richárd Nagyfi, and 1 others. 2024. Openassistant conversations-democratizing large language model alignment. Advances in Neural Information Processing Sys…
  21. Costas Mavromatis, Petros Karypis, and George Karypis. 2024. Pack of llms: Model fusion at test-time via perplexity optimization. Preprint, arXiv:2404.11531. — [arXiv:2404.11531](https://arxiv.org/abs/2404.11531) find
  22. Tommaso Mencattini, Adrian Robert Minut, Donato Crisostomi, Andrea Santilli, and Emanuele Rodolà. 2025. Merge 3 : Efficient evolutionary merging on consumer-grade gpus. arXiv preprint arXiv:2502.10436. find
  23. Subhabrata Mukherjee, Arindam Mitra, Ganesh Jawahar, Sahaj Agarwal, Hamid Palangi, and Ahmed Awadallah. 2023. Orca: Progressive learning from complex explanation traces of gpt-4. Preprint, arXiv:2306.02707.
  24. Mohammed Muqeeth, Haokun Liu, Yufan Liu, and Colin Raffel. 2024. Learning to route among specialized experts for zero-shot generalization. arXiv preprint arXiv:2402.05859. find
  25. Baolin Peng, Chunyuan Li, Pengcheng He, Michel Galley, and Jianfeng Gao. 2023. Instruction tuning with gpt-4. arXiv preprint arXiv:2304.03277.
  26. Soujanya Poria, Devamanyu Hazarika, Navonil Majumder, Gautam Naik, Erik Cambria, and Rada Mihalcea. 2018. Meld: A multimodal multi-party dataset for emotion recognition in conversations. arXiv preprint arXiv:1810.02508.
  27. Chengsong Huang, Qian Liu, Bill Yuchen Lin, Tianyu Pang, Chao Du, and Min Lin. 2024. Lorahub: Efficient cross-task generalization via dynamic lora composition. Preprint, arXiv:2307.13269.
  28. Qwen, :, An Yang, Baosong Yang, Beichen Zhang, Binyuan Hui, Bo Zheng, Bowen Yu, Chengyuan Li, Dayiheng Liu, Fei Huang, Haoran Wei, Huan Lin, Jian Yang, Jianhong Tu, Jianwei Zhang, Jianxin Yang, Jiaxi Yang, Jingren Zhou, and 25 others. 2025. Qwen2.5 technical report. Preprint, arXiv:2412.15115. find
  29. Hamish Ivison, Yizhong Wang, Valentina Pyatkin, Nathan Lambert, Matthew Peters, Pradeep Dasigi, Joel Jang, David Wadden, Noah A Smith, Iz Beltagy, and 1 others. 2023a. Camels in a changing climate: Enhancing lm adaptation with tulu 2. arXiv preprint arXiv:2311.10702.
  30. Freda Shi, Mirac Suzgun, Markus Freitag, Xuezhi Wang, Suraj Srivats, Soroush Vosoughi, Hyung Won Chung, Yi Tay, Sebastian Ruder, Denny Zhou, Dipanjan Das, and Jason Wei. 2022. Language models are multilingual chain-of-thought reasoners. Preprint, arXiv:2210.03057.
  31. Mirac Suzgun, Nathan Scales, Nathanael Schärli, Sebastian Gehrmann, Yi Tay, Hyung Won Chung, Aakanksha Chowdhery, Quoc V Le, Ed H Chi, Denny Zhou, , and Jason Wei. 2022. Challenging big-bench tasks and whether chain-of-thought can solve them. arXiv preprint arXiv:2210.09261.
  32. Alon Talmor, Jonathan Herzig, Nicholas Lourie, and Jonathan Berant. 2019. CommonsenseQA: A question answering challenge targeting commonsense knowledge. In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies,…
  33. Gemma Team, Morgane Riviere, Shreya Pathak, Pier Giuseppe Sessa, Cassidy Hardin, Surya Bhupatiraju, Léonard Hussenot, Thomas Mesnard, Bobak Shahriari, Alexandre Ramé, and 1 others. 2024. Gemma 2: Improving open language models at a practical size. arXiv preprint arXiv:2408.00118. find
  34. TechCrunch. 2023. Sharegpt website. find
  35. Junlin Wang, Jue Wang, Ben Athiwaratkun, Ce Zhang, and James Zou. 2024a. Mixture-of-agents enhances large language model capabilities. arXiv preprint arXiv:2406.04692. find
  36. Yubo Wang, Xueguang Ma, Ge Zhang, Yuansheng Ni, Abhranil Chandra, Shiguang Guo, Weiming Ren, Aaran Arulraj, Xuan He, Ziyan Jiang, and 1 others. 2024b. Mmlu-pro: A more robust and challenging multi-task language understanding benchmark. arXiv preprint arXiv:2406.01574.
  37. Mitchell Wortsman, Gabriel Ilharco, Samir Ya Gadre, Rebecca Roelofs, Raphael Gontijo-Lopes, Ari S Morcos, Hongseok Namkoong, Ali Farhadi, Yair Carmon, Simon Kornblith, and Ludwig Schmidt. 2022. Model soups: averaging weights of multiple fine-tuned models improves accuracy without increasing inferenc…
  38. Can Xu, Qingfeng Sun, Kai Zheng, Xiubo Geng, Pu Zhao, Jiazhan Feng, Chongyang Tao, and Daxin Jiang. 2023. Wizardlm: Empowering large language models to follow complex instructions. arXiv preprint arXiv:2304.12244.
  39. Prateek Yadav, Derek Tam, Leshem Choshen, Colin Raffel, and Mohit Bansal. 2023. Resolving interference when merging models. arXiv preprint arXiv:2306.01708, 1.
  40. Prateek Yadav, Tu Vu, Jonathan Lai, Alexandra Chronopoulou, Manaal Faruqui, Mohit Bansal, and Tsendsuren Munkhdalai. 2024. What matters for model merging at scale? arXiv preprint arXiv:2410.03617. find
  41. Enneng Yang, Zhenyi Wang, Li Shen, Shiwei Liu, Guibing Guo, Xingwei Wang, and Dacheng Tao. 2024. Adamerging: Adaptive model merging for multi-task learning. Preprint, arXiv:2310.02575. find
  42. Le Yu, Bowen Yu, Haiyang Yu, Fei Huang, and Yongbin Li. 2024. Language models are super mario: Absorbing abilities from homologous models as a free lunch. In Forty-first International Conference on Machine Learning.
  43. Sayyed M Zahiri and Jinho D Choi. 2018. Emotion detection on tv show transcripts with sequence-based convolutional neural networks. In Workshops at the thirty-second aaai conference on artificial intelligence. find
  44. Tianyi Zhang, Varsha Kishore, Felix Wu, Kilian Q Weinberger, and Yoav Artzi. 2019. Bertscore: Evaluating text generation with bert. arXiv preprint arXiv:1904.09675. find
  45. Ziyu Zhao, Leilei Gan, Guoyin Wang, Wangchunshu Zhou, Hongxia Yang, Kun Kuang, and Fei Wu. 2024. Loraretriever: Input-aware lora retrieval and composition for mixed tasks in the wild. Preprint, arXiv:2402.09997. find
  46. Yaowei Zheng, Richong Zhang, Junhao Zhang, Yanhan Ye, Zheyan Luo, Zhangchi Feng, and Yongqiang Ma. 2024. Llamafactory: Unified efficient fine-tuning of 100+ language models. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 3: System Demonstrations),…
  47. Chunting Zhou, Pengfei Liu, Puxin Xu, Srinivasan Iyer, Jiao Sun, Yuning Mao, Xuezhe Ma, Avia Efrat, Ping Yu, LILI YU, Susan Zhang, Gargi Ghosh, Mike Lewis, Luke Zettlemoyer, and Omer Levy. 2023. Lima: Less is more for alignment. In Advances in Neural Information Processing Systems, volume 36, pages… find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/zhang-2026-nature-inspired-population-based-evolution/