# .tmu: A Low-Entropy Tree-Structured Representation for LLM-Assisted Scientific Writing

arXiv: https://arxiv.org/abs/2603.02873 · HTML: https://mlsanity.com/paper/2603.02873/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 7 of its 31 references that are in the map.
- First posted on arXiv 3 Mar 2026.

## Builds on (most cited first, 7 of 7)

- [The AI Scientist: Towards Fully Automated Open-Ended Scientific Discovery](https://mlsanity.com/paper/2408.06292/) (2024)
- [Galactica: A Large Language Model for Science](https://mlsanity.com/paper/2211.09085/) (2022)
- [OpenWebMath: An Open Dataset of High-Quality Mathematical Web Text](https://mlsanity.com/paper/2310.06786/) (2023)
- [MarkupLM: Pre-training of Text and Markup Language for Visually-rich Document Understanding](https://mlsanity.com/paper/2110.08518/) (2021)
- [Table2LaTeX-RL: High-Fidelity LaTeX Code Generation from Table Images via Reinforced Multimodal Language Models](https://mlsanity.com/paper/2509.17589/) (2025)
- [TeXpert: A Multi-Level Benchmark for Evaluating LaTeX Code Generation by LLMs](https://mlsanity.com/paper/2506.16990/) (2025)
- [NeuRaLaTeX: A machine learning library written in pure LaTeX](https://mlsanity.com/paper/2503.24187/) (2025)

## Abstract (arXiv)

As large language models (LLMs) increasingly assist scientific writing, the limitations and token costs of generating TeX become increasingly visible. This paper analyzes TeX's architectural mismatch with LLM workflows, stemming from its lack of an explicit structural representation, to illustrate its limitations on generated semantics and error localization. As an alternative, we introduce .tmu, a low-entropy tree-structured representation. With its efficient data structure and clear contextual boundaries, .tmu outperforms .tex in the above aspects. Experiments across four LLMs provide evidence for this claim in most evaluated settings. Furthermore, we show that due to its lower information entropy, fine-tuning LLMs on .tmu achieves approximately 43% lower final training loss than on .tex. Our work provides a more scalable and LLM-friendly data representation for LLM-assisted scientific writing.