Elementary Math Word Problem Generation using Large Language Models

Elementary Math Word Problem Generation using Large Language Models | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Elementary Math Word Problem Generation using Large Language Models Nimesh Ariyarathne, Harshani Bandara, Yasith Heshan, Omega Gamage, and 8 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6814693/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Mathematics is often perceived as a complex subject by students, leading to high failure rates in exams. To improve Mathematics skills, it is important to provide sample questions for students to practice problem-solving. Manually creating Math Word Problems (MWPs) is time consuming for tutors, because they have to type in natural language while adhering to grammar and spelling rules of the language. Existing Deep Learning techniques for MWP generation either require a tutor to provide the initial portion of the MWP, and/or additional information such as an equation. In this paper, we present an MWP generation system based on Large Language Models (LLMs) that overcome the need for additional input - the only input to our system is the number of MWPs needed, the grade and the type of question (e.g.~addition, subtraction). Unlike the existing LLM-based solutions for MWP generation, we carried out an extensive set of experiments involving different LLMs, prompting strategies, techniques to improve the diversity of questions, as well as techniques that employ human feedback to improve LLM performance. Human and automated evaluations confirmed that the generated MWPs are high in quality, with minimal spelling and grammar issues. However, LLMs still struggle to generate questions that adhere to the specified grade and question type requirements. Full Text Additional Declarations No competing interests reported. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6814693","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":507922734,"identity":"88e5d6ba-9282-4dc3-ad40-b7892e596bf8","order_by":0,"name":"Nimesh Ariyarathne","email":"","orcid":"","institution":"University of Moratuwa","correspondingAuthor":false,"prefix":"","firstName":"Nimesh","middleName":"","lastName":"Ariyarathne","suffix":""},{"id":507922735,"identity":"03978085-4c96-4bc4-8b95-eb7d9e9eaee2","order_by":1,"name":"Harshani Bandara","email":"","orcid":"","institution":"University of Moratuwa","correspondingAuthor":false,"prefix":"","firstName":"Harshani","middleName":"","lastName":"Bandara","suffix":""},{"id":507922736,"identity":"55a371d2-e576-4065-ac49-5e57ff28e435","order_by":2,"name":"Yasith Heshan","email":"","orcid":"","institution":"University of Moratuwa","correspondingAuthor":false,"prefix":"","firstName":"Yasith","middleName":"","lastName":"Heshan","suffix":""},{"id":507922737,"identity":"274ecfcf-eb02-4d80-969d-a42be1a53f55","order_by":3,"name":"Omega Gamage","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Omega","middleName":"","lastName":"Gamage","suffix":""},{"id":507922738,"identity":"d7671814-c409-48ce-b6c2-1c322b1c6e6d","order_by":4,"name":"Surangika Ranathunga","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABDUlEQVRIiWNgGAWjYBCDBDD5gQQdBmAtjDNI1sLMQ4xa+Qbeh58Lav7kGdxuv/jZNsfOXr69x4DhRw1DnnwDDuMPsBtLzzhmUGxw50yxdO625MQNZ84YMPYcYyhmxKWFgY1BmofNIHHDjZwEoBbmBAOJHAMG3gaGxGacDmNj/s3zD6wl+bfltnp7+flvDBj/ArW04fLMATY2ad42kJb0Y9KM2w4zNtzgMWAG2dKDQ4fBYTY2a94+48SZN3LYLHu3HQf6Ja3gsMwxiURcIS7f3sZ8m+ebXGLfjfTHN35uqwaG2OGND9/U2CTOx+F9BmY4i8cAyuAwOMDAIIHLJ8iA/QE6YxSMglEwCkYBGAAAZ+NYcU4dsWEAAAAASUVORK5CYII=","orcid":"","institution":"Massey University","correspondingAuthor":true,"prefix":"","firstName":"Surangika","middleName":"","lastName":"Ranathunga","suffix":""},{"id":507922740,"identity":"9b99b2f3-d7c3-4f67-9823-dea82103bd81","order_by":5,"name":"Dilan Nayanajith","email":"","orcid":"","institution":"University of Moratuwa","correspondingAuthor":false,"prefix":"","firstName":"Dilan","middleName":"","lastName":"Nayanajith","suffix":""},{"id":507922742,"identity":"cfa04754-0db2-4efc-b501-44d7b005c5f4","order_by":6,"name":"Yutharsan Sivapalan","email":"","orcid":"","institution":"University of Moratuwa","correspondingAuthor":false,"prefix":"","firstName":"Yutharsan","middleName":"","lastName":"Sivapalan","suffix":""},{"id":507922744,"identity":"e5223b5a-5b55-45e8-95a4-4c8a5a800aea","order_by":7,"name":"Gayathri Lihinikaduarachchi","email":"","orcid":"","institution":"University of Moratuwa","correspondingAuthor":false,"prefix":"","firstName":"Gayathri","middleName":"","lastName":"Lihinikaduarachchi","suffix":""},{"id":507922746,"identity":"a45a0f28-cd69-4477-9661-a12529d4db86","order_by":8,"name":"Tharoosha Vihidun","email":"","orcid":"","institution":"University of Moratuwa","correspondingAuthor":false,"prefix":"","firstName":"Tharoosha","middleName":"","lastName":"Vihidun","suffix":""},{"id":507922748,"identity":"68f39614-18b7-4f5b-aea9-937bd8a4acdd","order_by":9,"name":"Meenambika Chandirakumar","email":"","orcid":"","institution":"University of Moratuwa","correspondingAuthor":false,"prefix":"","firstName":"Meenambika","middleName":"","lastName":"Chandirakumar","suffix":""},{"id":507922750,"identity":"bc948402-f53f-4737-82d9-5ecbb059c152","order_by":10,"name":"Sanujen Premakumar","email":"","orcid":"","institution":"University of Moratuwa","correspondingAuthor":false,"prefix":"","firstName":"Sanujen","middleName":"","lastName":"Premakumar","suffix":""},{"id":507922752,"identity":"d98bddf3-5f29-4522-b49c-b86c4e31ce95","order_by":11,"name":"Sanjula Gathsara","email":"","orcid":"","institution":"University of Moratuwa","correspondingAuthor":false,"prefix":"","firstName":"Sanjula","middleName":"","lastName":"Gathsara","suffix":""}],"badges":[],"createdAt":"2025-06-03 22:23:05","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6814693/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6814693/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":104964148,"identity":"ac503621-942f-4a97-a866-2f20474cc113","added_by":"auto","created_at":"2026-03-19 09:28:47","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1570354,"visible":true,"origin":"","legend":"","description":"","filename":"Mathwordproblemgenerationmathwizardspaperfinal1.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6814693/v1_covered_9988af8b-b3b8-468c-bc9a-b578cfa2a71e.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Elementary Math Word Problem Generation using Large Language Models","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":true,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":true,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-6814693/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6814693/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"Mathematics is often perceived as a complex subject by students, leading to high failure rates in exams. To improve Mathematics skills, it is important to provide sample questions for students to practice problem-solving. Manually creating Math Word Problems (MWPs) is time consuming for tutors, because they have to type in natural language while adhering to grammar and spelling rules of the language. Existing Deep Learning techniques for MWP generation either require a tutor to provide the initial portion of the MWP, and/or additional information such as an equation. In this paper, we present an MWP generation system based on Large Language Models (LLMs) that overcome the need for additional input - the only input to our system is the number of MWPs needed, the grade and the type of question (e.g.~addition, subtraction). Unlike the existing LLM-based solutions for MWP generation, we carried out an extensive set of experiments involving different LLMs, prompting strategies, techniques to improve the diversity of questions, as well as techniques that employ human feedback to improve LLM performance. Human and automated evaluations confirmed that the generated MWPs are high in quality, with minimal spelling and grammar issues. However, LLMs still struggle to generate questions that adhere to the specified grade and question type requirements.","manuscriptTitle":"Elementary Math Word Problem Generation using Large Language Models","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-09-02 04:44:43","doi":"10.21203/rs.3.rs-6814693/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"1631dc12-104b-4d07-845c-a57faf45397f","owner":[],"postedDate":"September 2nd, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2026-03-19T09:27:05+00:00","versionOfRecord":[],"versionCreatedAt":"2025-09-02 04:44:43","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6814693","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6814693","identity":"rs-6814693","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}