https://arxiv.org/abs/2512.24617 Skip to main content Cornell University We gratefully acknowledge support from the Simons Foundation, member institutions, and all contributors. Donate arxiv logo > cs > arXiv:2512.24617 [ ] Help | Advanced Search [All fields ] Search arXiv logo Cornell University Logo [ ] GO quick links * Login * Help Pages * About Computer Science > Machine Learning arXiv:2512.24617 (cs) [Submitted on 31 Dec 2025 (v1), last revised 5 Jan 2026 (this version, v2)] Title:Dynamic Large Concept Models: Latent Reasoning in an Adaptive Semantic Space Authors:Xingwei Qu, Shaowen Wang, Zihao Huang, Kai Hua, Fan Yin, Rui-Jie Zhu, Jundong Zhou, Qiyang Min, Zihao Wang, Yizhi Li, Tianyu Zhang, He Xing, Zheng Zhang, Yuxuan Song, Tianyu Zheng, Zhiyuan Zeng, Chenghua Lin, Ge Zhang, Wenhao Huang View a PDF of the paper titled Dynamic Large Concept Models: Latent Reasoning in an Adaptive Semantic Space, by Xingwei Qu and 18 other authors View PDF HTML (experimental) Abstract:Large Language Models (LLMs) apply uniform computation to all tokens, despite language exhibiting highly non-uniform information density. This token-uniform regime wastes capacity on locally predictable spans while under-allocating computation to semantically critical transitions. We propose $\textbf{Dynamic Large Concept Models (DLCM)}$, a hierarchical language modeling framework that learns semantic boundaries from latent representations and shifts computation from tokens to a compressed concept space where reasoning is more efficient. DLCM discovers variable-length concepts end-to-end without relying on predefined linguistic units. Hierarchical compression fundamentally changes scaling behavior. We introduce the first $\ textbf{compression-aware scaling law}$, which disentangles token-level capacity, concept-level reasoning capacity, and compression ratio, enabling principled compute allocation under fixed FLOPs. To stably train this heterogeneous architecture, we further develop a $\textbf{decoupled $\mu$P parametrization}$ that supports zero-shot hyperparameter transfer across widths and compression regimes. At a practical setting ($R=4$, corresponding to an average of four tokens per concept), DLCM reallocates roughly one-third of inference compute into a higher-capacity reasoning backbone, achieving a $\textbf{+2.69$\%$ average improvement}$ across 12 zero-shot benchmarks under matched inference FLOPs. Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI) Cite as: arXiv:2512.24617 [cs.LG] (or arXiv:2512.24617v2 [cs.LG] for this version) https://doi.org/10.48550/arXiv.2512.24617 Focus to learn more arXiv-issued DOI via DataCite Submission history From: Xingwei Qu [view email] [v1] Wed, 31 Dec 2025 04:19:33 UTC (2,886 KB) [v2] Mon, 5 Jan 2026 05:44:29 UTC (2,887 KB) Full-text links: Access Paper: View a PDF of the paper titled Dynamic Large Concept Models: Latent Reasoning in an Adaptive Semantic Space, by Xingwei Qu and 18 other authors * View PDF * HTML (experimental) * TeX Source license icon view license Current browse context: cs.LG < prev | next > new | recent | 2025-12 Change to browse by: cs cs.AI References & Citations * NASA ADS * Google Scholar * Semantic Scholar export BibTeX citation Loading... BibTeX formatted citation x [loading... ] Data provided by: Bookmark BibSonomy logo Reddit logo (*) Bibliographic Tools Bibliographic and Citation Tools [ ] Bibliographic Explorer Toggle Bibliographic Explorer (What is the Explorer?) [ ] Connected Papers Toggle Connected Papers (What is Connected Papers?) [ ] Litmaps Toggle Litmaps (What is Litmaps?) [ ] scite.ai Toggle scite Smart Citations (What are Smart Citations?) ( ) Code, Data, Media Code, Data and Media Associated with this Article [ ] alphaXiv Toggle alphaXiv (What is alphaXiv?) [ ] Links to Code Toggle CatalyzeX Code Finder for Papers (What is CatalyzeX?) [ ] DagsHub Toggle DagsHub (What is DagsHub?) [ ] GotitPub Toggle Gotit.pub (What is GotitPub?) [ ] Huggingface Toggle Hugging Face (What is Huggingface?) [ ] Links to Code Toggle Papers with Code (What is Papers with Code?) [ ] ScienceCast Toggle ScienceCast (What is ScienceCast?) ( ) Demos Demos [ ] Replicate Toggle Replicate (What is Replicate?) [ ] Spaces Toggle Hugging Face Spaces (What is Spaces?) [ ] Spaces Toggle TXYZ.AI (What is TXYZ.AI?) ( ) Related Papers Recommenders and Search Tools [ ] Link to Influence Flower Influence Flower (What are Influence Flowers?) [ ] Core recommender toggle CORE Recommender (What is CORE?) [ ] IArxiv recommender toggle IArxiv Recommender (What is IArxiv?) * Author * Venue * Institution * Topic ( ) About arXivLabs arXivLabs: experimental projects with community collaborators arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website. Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them. Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs. Which authors of this paper are endorsers? | Disable MathJax (What is MathJax?) * About * Help * Click here to contact arXiv Contact * Click here to subscribe Subscribe * Copyright * Privacy Policy * Web Accessibility Assistance * arXiv Operational Status