LLM DNA: Tracing Model Evolution via Functional Representations

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Wu, Zhaomin, Zhao, Haodong, Wang, Ziyang, Guo, Jizhou, Wang, Qian, He, Bingsheng
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915971308978176
author Wu, Zhaomin
Zhao, Haodong
Wang, Ziyang
Guo, Jizhou
Wang, Qian
He, Bingsheng
author_facet Wu, Zhaomin
Zhao, Haodong
Wang, Ziyang
Guo, Jizhou
Wang, Qian
He, Bingsheng
contents The explosive growth of large language models (LLMs) has created a vast but opaque landscape: millions of models exist, yet their evolutionary relationships through fine-tuning, distillation, or adaptation are often undocumented or unclear, complicating LLM management. Existing methods are limited by task specificity, fixed model sets, or strict assumptions about tokenizers or architectures. Inspired by biological DNA, we address these limitations by mathematically defining LLM DNA as a low-dimensional, bi-Lipschitz representation of functional behavior. We prove that LLM DNA satisfies inheritance and genetic determinism properties and establish the existence of DNA. Building on this theory, we derive a general, scalable, training-free pipeline for DNA extraction. In experiments across 305 LLMs, DNA aligns with prior studies on limited subsets and achieves superior or competitive performance on specific tasks. Beyond these tasks, DNA comparisons uncover previously undocumented relationships among LLMs. We further construct the evolutionary tree of LLMs using phylogenetic algorithms, which align with shifts from encoder-decoder to decoder-only architectures, reflect temporal progression, and reveal distinct evolutionary speeds across LLM families.
format Preprint
id arxiv_https___arxiv_org_abs_2509_24496
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle LLM DNA: Tracing Model Evolution via Functional Representations
Wu, Zhaomin
Zhao, Haodong
Wang, Ziyang
Guo, Jizhou
Wang, Qian
He, Bingsheng
Machine Learning
Artificial Intelligence
The explosive growth of large language models (LLMs) has created a vast but opaque landscape: millions of models exist, yet their evolutionary relationships through fine-tuning, distillation, or adaptation are often undocumented or unclear, complicating LLM management. Existing methods are limited by task specificity, fixed model sets, or strict assumptions about tokenizers or architectures. Inspired by biological DNA, we address these limitations by mathematically defining LLM DNA as a low-dimensional, bi-Lipschitz representation of functional behavior. We prove that LLM DNA satisfies inheritance and genetic determinism properties and establish the existence of DNA. Building on this theory, we derive a general, scalable, training-free pipeline for DNA extraction. In experiments across 305 LLMs, DNA aligns with prior studies on limited subsets and achieves superior or competitive performance on specific tasks. Beyond these tasks, DNA comparisons uncover previously undocumented relationships among LLMs. We further construct the evolutionary tree of LLMs using phylogenetic algorithms, which align with shifts from encoder-decoder to decoder-only architectures, reflect temporal progression, and reveal distinct evolutionary speeds across LLM families.
title LLM DNA: Tracing Model Evolution via Functional Representations
topic Machine Learning
Artificial Intelligence
url https://arxiv.org/abs/2509.24496