A LoRA-Based Approach to Fine-Tuning LLMs for Educational Guidance in Resource-Constrained Settings

A LoRA-Based Approach to Fine-Tuning LLMs for Educational Guidance in Resource-Constrained Settings | 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 A LoRA-Based Approach to Fine-Tuning LLMs for Educational Guidance in Resource-Constrained Settings MD MILLAT HOSEN This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8431533/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 The current study describes a cost-effective method for adapting large language models(LLMs) for academic advising in study-abroad contexts. Using the Mistral-7B model withLow-Rank Adaptation (LoRA) and 4-bit NF4 quantization via the Unsloth framework, themodel underwent training in two distinct hardware phases to demonstrate adaptability andcomputational efficiency. Contrary to multi-stage data curation, this study utilized a sin-gle synthetic dataset of 2,274 conversation pairs across both phases to evaluate hardware-agnostic convergence. In Phase 1, the model was fine-tuned on an NVIDIA Tesla P100.In Phase 2, the model continued training on the same dataset using an NVIDIA Tesla T4with optimized batch configurations to refine performance. Technical innovations utilizedmemory-efficient quantization and continuous training analytics. After training, the studydemonstrated a total reduction in training loss from ∼1.01 to ∼0.34, achieving stable conver-gence on consumer-grade GPU equipment. These findings support the effective applicationof instruction-tuned LLMs within educational advising, specifically showing that trainingcan be effectively distributed and resumed across varying hardware architectures. 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-8431533","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":564661732,"identity":"bf14c625-686e-40e4-9282-820737d5ea88","order_by":0,"name":"MD MILLAT HOSEN","email":"data:image/png;base64,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","orcid":"","institution":"Sharda University","correspondingAuthor":true,"prefix":"","firstName":"MD","middleName":"MILLAT","lastName":"HOSEN","suffix":""}],"badges":[],"createdAt":"2025-12-23 07:53:45","currentVersionCode":1,"declarations":{"humanSubjects":false,"vertebrateSubjects":false,"conflictsOfInterestStatement":false,"humanSubjectEthicalGuidelines":false,"humanSubjectConsent":false,"humanSubjectClinicalTrial":false,"humanSubjectCaseReport":false,"vertebrateSubjectEthicalGuidelines":false},"doi":"10.21203/rs.3.rs-8431533/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8431533/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":98987805,"identity":"7d23d310-49ea-43ce-a374-b3b391ceb2bf","added_by":"auto","created_at":"2025-12-25 09:18:17","extension":"json","order_by":0,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":3664,"visible":true,"origin":"","legend":"","description":"","filename":"b045225abb9141e49695d3aa8fe3a8f6.json","url":"https://assets-eu.researchsquare.com/files/rs-8431533/v1/7b13a656b8879dbc347bf5ba.json"},{"id":98987804,"identity":"c48d0a50-5429-450a-ae3f-83d85f3a0451","added_by":"auto","created_at":"2025-12-25 09:18:17","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":32588,"visible":true,"origin":"","legend":"","description":"","filename":"Coverletter.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8431533/v1/ce42da0f156b51c4f3792132.pdf"},{"id":99311924,"identity":"7f9c79cc-1f67-4fbf-a1c3-ce395720bc67","added_by":"auto","created_at":"2025-12-31 16:17:22","extension":"pdf","order_by":2,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":6323524,"visible":true,"origin":"","legend":"","description":"","filename":"StudyAbroadGPTSourceFilesAnonymized.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8431533/v1/97f159626e29ffe1b6996c58.pdf"},{"id":98987811,"identity":"1f0bbd1b-eb2d-4a5b-81f1-c7a150ad7fb6","added_by":"auto","created_at":"2025-12-25 09:18:17","extension":"png","order_by":3,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":5475046,"visible":true,"origin":"","legend":"","description":"","filename":"fig1arch.png","url":"https://assets-eu.researchsquare.com/files/rs-8431533/v1/412f6ad90478815768c109d9.png"},{"id":98987806,"identity":"0dcca0ca-3108-4aba-8f85-b56255a43bf4","added_by":"auto","created_at":"2025-12-25 09:18:17","extension":"png","order_by":4,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":491999,"visible":true,"origin":"","legend":"","description":"","filename":"fig2lossp100.png","url":"https://assets-eu.researchsquare.com/files/rs-8431533/v1/4286a627c41b1a462fb1aaa5.png"},{"id":98987808,"identity":"0749622f-f863-489b-9ec5-cea8a79ffce0","added_by":"auto","created_at":"2025-12-25 09:18:17","extension":"png","order_by":5,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":457135,"visible":true,"origin":"","legend":"","description":"","filename":"fig3gradp100.png","url":"https://assets-eu.researchsquare.com/files/rs-8431533/v1/33ba7fc507e7a55a5c0d923f.png"},{"id":99312898,"identity":"ecc9cca3-1936-4ade-b0dc-5d3249516adb","added_by":"auto","created_at":"2025-12-31 16:19:34","extension":"png","order_by":6,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":454616,"visible":true,"origin":"","legend":"","description":"","filename":"fig4losst4.png","url":"https://assets-eu.researchsquare.com/files/rs-8431533/v1/14616d11ef1653bec9b186e6.png"},{"id":98987807,"identity":"57bef5b5-98c0-455a-9d29-8215cab41ab8","added_by":"auto","created_at":"2025-12-25 09:18:17","extension":"png","order_by":7,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":438257,"visible":true,"origin":"","legend":"","description":"","filename":"fig5gradt4.png","url":"https://assets-eu.researchsquare.com/files/rs-8431533/v1/5b2b2de66d4b78fc0e230111.png"},{"id":98987812,"identity":"18058ab5-3068-44db-8e23-8a608b1b13a3","added_by":"auto","created_at":"2025-12-25 09:18:17","extension":"xml","order_by":8,"title":"","display":"","copyAsset":false,"role":"acdc-reference","size":24073,"visible":true,"origin":"","legend":"","description":"","filename":"b045225abb9141e49695d3aa8fe3a8f61structuring.xml","url":"https://assets-eu.researchsquare.com/files/rs-8431533/v1/e49aebf47a3d257c4fcad51d.xml"},{"id":99787875,"identity":"6d9def51-de5d-4d02-bfae-85cec2ddd6bc","added_by":"auto","created_at":"2026-01-08 12:40:27","extension":"pdf","order_by":1,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1397577,"visible":true,"origin":"","legend":"","description":"","filename":"StudyAbroadGPTSourceFilesAnonymized.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8431533/v1_covered_60fecccf-3576-4d2f-b9cb-00f0ce828c39.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"A LoRA-Based Approach to Fine-Tuning LLMs for Educational Guidance in Resource-Constrained Settings","fulltext":[],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":false,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"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-8431533/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8431533/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"The current study describes a cost-effective method for adapting large language models(LLMs) for academic advising in study-abroad contexts. Using the Mistral-7B model withLow-Rank Adaptation (LoRA) and 4-bit NF4 quantization via the Unsloth framework, themodel underwent training in two distinct hardware phases to demonstrate adaptability andcomputational efficiency. Contrary to multi-stage data curation, this study utilized a sin-gle synthetic dataset of 2,274 conversation pairs across both phases to evaluate hardware-agnostic convergence. In Phase 1, the model was fine-tuned on an NVIDIA Tesla P100.In Phase 2, the model continued training on the same dataset using an NVIDIA Tesla T4with optimized batch configurations to refine performance. Technical innovations utilizedmemory-efficient quantization and continuous training analytics. After training, the studydemonstrated a total reduction in training loss from ∼1.01 to ∼0.34, achieving stable conver-gence on consumer-grade GPU equipment. These findings support the effective applicationof instruction-tuned LLMs within educational advising, specifically showing that trainingcan be effectively distributed and resumed across varying hardware architectures.","manuscriptTitle":"A LoRA-Based Approach to Fine-Tuning LLMs for Educational Guidance in Resource-Constrained Settings","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-12-25 09:18:12","doi":"10.21203/rs.3.rs-8431533/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":"1f28120f-3ea3-4c51-a076-2076114a6948","owner":[],"postedDate":"December 25th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2025-12-29T10:24:31+00:00","versionOfRecord":[],"versionCreatedAt":"2025-12-25 09:18:12","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8431533","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8431533","identity":"rs-8431533","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}