{"slug": "fractal-basins-trap-latent-reasoning", "title": "Fractal basins trap latent reasoning", "summary": "A study submitted to arXiv on 4 Sep 2026 by William Gilpin finds that leading AI reasoning models exhibit transient chaos and fractal basins, with fractality increasing with task difficulty across tasks like Sudoku, maze solving, visual puzzles, and mathematical logic. The research shows reasoning slowdowns are an inevitable consequence of problem hardness, as models become trapped near saddle points corresponding to nearly-correct solutions.", "body_md": "# Computer Science > Machine Learning\n\n  [Submitted on 4 Sep 2026]\n\n# Title:Fractal basins trap latent reasoning\n\n[View PDF](/pdf/2609.04963)\n\n[HTML (experimental)](https://arxiv.org/html/2609.04963v1)\n\nAbstract:Reasoning allows artificial intelligence models to revisit and correct their mistakes, enabling recent frontier advances in mathematical theorem solving, software engineering, and autonomous task planning. Reasoning models are widely observed to reason for longer on harder tasks, but the general mechanism responsible for these slowdowns is unknown. Here, we show that reasoning models exhibit transient chaos, a physical consequence of the computational complexity of difficult tasks. As a consequence, we show that diverse leading reasoning models are dynamical systems with fractal basins, with fractality increasing with task difficulty across diverse tasks like Sudoku and maze solving, visual puzzles, and mathematical logic. We show that transient chaos emerges due to reasoning becoming trapped for extended durations near saddle points, which we show correspond to nearly-correct attempted solutions of the underlying problem. Our results show that reasoning slowdowns are an inevitable consequence of problem hardness in modern artificial intelligence models, and establish reasoning traces as a rich new class of dynamical system.\n    \n\n## Submission history\n\nFrom: William Gilpin [\n[view email](/show-email/7195b41d/2609.04963)] [via William Gilpin as proxy]\n\n**[v1]** Fri, 4 Sep 2026 10:11:36 UTC (17,726 KB)\n\n### References & Citations\n\nLoading...\n\n# Bibliographic and Citation Tools\n\nBibliographic Explorer \n\n*(*[What is the Explorer?](https://info.arxiv.org/labs/showcase.html#arxiv-bibliographic-explorer))\nConnected Papers \n\n*(*[What is Connected Papers?](https://www.connectedpapers.com/about))\nLitmaps \n\n*(*[What is Litmaps?](https://www.litmaps.co/))\nscite Smart Citations \n\n*(*[What are Smart Citations?](https://www.scite.ai/))\n# Code, Data and Media Associated with this Article\n\nalphaXiv \n\n*(*[What is alphaXiv?](https://alphaxiv.org/))\nCatalyzeX Code Finder for Papers \n\n*(*[What is CatalyzeX?](https://www.catalyzex.com))\nDagsHub \n\n*(*[What is DagsHub?](https://dagshub.com/))\nGotit.pub \n\n*(*[What is GotitPub?](http://gotit.pub/faq))\nHugging Face \n\n*(*[What is Huggingface?](https://huggingface.co/huggingface))\nScienceCast \n\n*(*[What is ScienceCast?](https://sciencecast.org/welcome))\n# Demos\n\n# Recommenders and Search Tools\n\nInfluence Flower \n\n*(*[What are Influence Flowers?](https://influencemap.cmlab.dev/))\nCORE Recommender \n\n*(*[What is CORE?](https://core.ac.uk/services/recommender))\nIArxiv Recommender\n\n*(*[What is IArxiv?](https://iarxiv.org/about))\n# arXivLabs: experimental projects with community collaborators\n\narXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.\n\nBoth 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.\n\nHave an idea for a project that will add value for arXiv's community? [**Learn more about arXivLabs**](https://info.arxiv.org/labs/index.html).", "url": "https://wpnews.pro/news/fractal-basins-trap-latent-reasoning", "canonical_source": "https://arxiv.org/abs/2609.04963", "published_at": "2026-09-08 19:07:38+00:00", "updated_at": "2026-09-08 19:19:23.878016+00:00", "lang": "en", "topics": ["machine-learning", "artificial-intelligence", "ai-research"], "entities": ["William Gilpin", "arXiv"], "alternates": {"html": "https://wpnews.pro/news/fractal-basins-trap-latent-reasoning", "markdown": "https://wpnews.pro/news/fractal-basins-trap-latent-reasoning.md", "text": "https://wpnews.pro/news/fractal-basins-trap-latent-reasoning.txt", "jsonld": "https://wpnews.pro/news/fractal-basins-trap-latent-reasoning.jsonld"}}