CoinMath: Harnessing the Power of Coding Instruction for Math LLMs

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Wei, Chengwei, Wang, Bin, Kim, Jung-jae, Liu, Guimei, Chen, Nancy F.
Natura: Preprint
Pubblicazione: 2024
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866910747666153472
author Wei, Chengwei
Wang, Bin
Kim, Jung-jae
Liu, Guimei
Chen, Nancy F.
author_facet Wei, Chengwei
Wang, Bin
Kim, Jung-jae
Liu, Guimei
Chen, Nancy F.
contents Large Language Models (LLMs) have shown strong performance in solving mathematical problems, with code-based solutions proving particularly effective. However, the best practice to leverage coding instruction data to enhance mathematical reasoning remains underexplored. This study investigates three key questions: (1) How do different coding styles of mathematical code-based rationales impact LLMs' learning performance? (2) Can general-domain coding instructions improve performance? (3) How does integrating textual rationales with code-based ones during training enhance mathematical reasoning abilities? Our findings reveal that code-based rationales with concise comments, descriptive naming, and hardcoded solutions are beneficial, while improvements from general-domain coding instructions and textual rationales are relatively minor. Based on these insights, we propose CoinMath, a learning strategy designed to enhance mathematical reasoning by diversifying the coding styles of code-based rationales. CoinMath generates a variety of code-based rationales incorporating concise comments, descriptive naming conventions, and hardcoded solutions. Experimental results demonstrate that CoinMath significantly outperforms its baseline model, MAmmoTH, one of the SOTA math LLMs.
format Preprint
id arxiv_https___arxiv_org_abs_2412_11699
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle CoinMath: Harnessing the Power of Coding Instruction for Math LLMs
Wei, Chengwei
Wang, Bin
Kim, Jung-jae
Liu, Guimei
Chen, Nancy F.
Computation and Language
Large Language Models (LLMs) have shown strong performance in solving mathematical problems, with code-based solutions proving particularly effective. However, the best practice to leverage coding instruction data to enhance mathematical reasoning remains underexplored. This study investigates three key questions: (1) How do different coding styles of mathematical code-based rationales impact LLMs' learning performance? (2) Can general-domain coding instructions improve performance? (3) How does integrating textual rationales with code-based ones during training enhance mathematical reasoning abilities? Our findings reveal that code-based rationales with concise comments, descriptive naming, and hardcoded solutions are beneficial, while improvements from general-domain coding instructions and textual rationales are relatively minor. Based on these insights, we propose CoinMath, a learning strategy designed to enhance mathematical reasoning by diversifying the coding styles of code-based rationales. CoinMath generates a variety of code-based rationales incorporating concise comments, descriptive naming conventions, and hardcoded solutions. Experimental results demonstrate that CoinMath significantly outperforms its baseline model, MAmmoTH, one of the SOTA math LLMs.
title CoinMath: Harnessing the Power of Coding Instruction for Math LLMs
topic Computation and Language
url https://arxiv.org/abs/2412.11699