English

CoMAT: Chain of Mathematically Annotated Thought Improves Mathematical Reasoning

Artificial Intelligence 2026-01-16 v2 Computation and Language Machine Learning Symbolic Computation

Abstract

Mathematical reasoning remains a significant challenge for large language models (LLMs), despite progress in prompting techniques such as Chain-of-Thought (CoT). We present **Chain of Mathematically Annotated Thought (CoMAT)**, which enhances reasoning through two stages: *Symbolic Conversion* (converting natural language queries into symbolic form) and *Reasoning Execution* (deriving answers from symbolic representations). CoMAT operates entirely with a single LLM and without external solvers. Across four LLMs, CoMAT outperforms traditional CoT on six out of seven benchmarks, achieving gains of 4.48% on MMLU-Redux (MATH) and 4.58% on GaoKao MCQ. In addition to improved performance, CoMAT ensures faithfulness and verifiability, offering a transparent reasoning process for complex mathematical tasks

Keywords

Cite

@article{arxiv.2410.10336,
  title  = {CoMAT: Chain of Mathematically Annotated Thought Improves Mathematical Reasoning},
  author = {Joshua Ong Jun Leang and Aryo Pradipta Gema and Shay B. Cohen},
  journal= {arXiv preprint arXiv:2410.10336},
  year   = {2026}
}

Comments

9 pages, 12 figures

R2 v1 2026-06-28T19:20:19.611Z