prikaz prve stranice dokumenta Primjena i usporedba LLM-ova u evaluaciji kôda
  Download
PDF 1.35 MB
master's thesis
Primjena i usporedba LLM-ova u evaluaciji kôda
2024. urn:nbn:hr:166:838963
Mrčela, Judita
University of Split
Faculty of Science

Cite this document

Mrčela, J. (2024). Primjena i usporedba LLM-ova u evaluaciji kôda (Master's thesis). Split: University of Split, Faculty of Science. Retrieved from https://urn.nsk.hr/urn:nbn:hr:166:838963

Mrčela, Judita. "Primjena i usporedba LLM-ova u evaluaciji kôda." Master's thesis, University of Split, Faculty of Science, 2024. https://urn.nsk.hr/urn:nbn:hr:166:838963

Mrčela, Judita. "Primjena i usporedba LLM-ova u evaluaciji kôda." Master's thesis, University of Split, Faculty of Science, 2024. https://urn.nsk.hr/urn:nbn:hr:166:838963

Mrčela, J. (2024). 'Primjena i usporedba LLM-ova u evaluaciji kôda', Master's thesis, University of Split, Faculty of Science, accessed 04 December 2024, https://urn.nsk.hr/urn:nbn:hr:166:838963

Mrčela J. Primjena i usporedba LLM-ova u evaluaciji kôda [Master's thesis]. Split: University of Split, Faculty of Science; 2024 [cited 2024 December 04] Available at: https://urn.nsk.hr/urn:nbn:hr:166:838963

J. Mrčela, "Primjena i usporedba LLM-ova u evaluaciji kôda", Master's thesis, University of Split, Faculty of Science, Split, 2024. Available at: https://urn.nsk.hr/urn:nbn:hr:166:838963

Please login to the repository to save this object to your list.
Metadata
TitlePrimjena i usporedba LLM-ova u evaluaciji kôda
Title (english)Application and Comparison of LLMs in Code Evaluation
AuthorJudita Mrčela
MentorGoran Zaharija (mentor)
Committee memberGoran Zaharija (predsjednik povjerenstva)
Committee memberDivna Krpan (član povjerenstva)
Committee memberMonika Mladenović (član povjerenstva)
GranterUniversity of Split
Faculty of Science
Split
Defense date and country2024-09-25, Croatia
Scientific / art field,
discipline and subdiscipline		TECHNICAL SCIENCES
Computing
Abstract
Ovo istraživanje za cilj ima istražiti primjenu i usporedbu velikih jezičnih modela u evaluaciji procjene koda. Usredotočeno je na tri modela: OpenAI-jev GPT-4, Code Llama 7B Instruct i fino podešenu verziju CodeT5+. Modeli se ocjenjuju na temelju njihove sposobnosti procjene koda u različitim kategorijama, uključujući matematičke i algoritamske zadatke, strukture podataka i web programiranje. Uspoređujući izvedbu ovih modela, nastoje se identificirati njihove prednosti i slabosti u pružanju točnih i korisnih povratnih informacija za poboljšanje kvalitete koda. Ovo istraživanje naglašava potencijal velikih jezičnih u pojednostavljivanju procesa ocjenjivanja programskog koda, a istovremeno prepoznaje potrebu za daljnjim napretkom u njihovom finom podešavanju i primjeni.
Abstract (english)
This research aims to investigate the application and comparison of large language models in the evaluation of code estimation. It focuses on three models: OpenAI's GPT-4, Code Llama 7B Instruct, and a fine-tuned version of CodeT5+. Models are evaluated based on their ability to evaluate code in a variety of categories, including mathematical and algorithmic tasks, data structures, and web programming. By comparing the performance of these models, we seek to identify their strengths and weaknesses in providing accurate and useful feedback to improve code quality. This research highlights the potential of large languages in simplifying the process of evaluating software code, while recognizing the need for further progress in their fine-tuning and application.
Keywords
Keywords (english)
Languagecroatian
URN:NBNurn:nbn:hr:166:838963
Study programmeTitle: Computer Science; specializations in: specialization in Education
Course: specialization in Education
Study programme type: university
Study level: graduate
Academic / professional title: sveučilišni/a magistar/magistra edukacije informatike (sveučilišni/a magistar/magistra edukacije informatike)
Type of resourceText
File originBorn digital
Access conditionsOpen access
Terms of useLicence In copyright icon
Created on2024-10-16 08:32:36