Tehnike lanca misli za rješavanje strateških problema

undergraduate thesis

• Title (english): Chain Of Thought Techniques For Solving Strategic Problems
• Author: Iva Jurković
• Mentor: Mario Brčić (mentor)
• Committee member: Mario Brčić (predsjednik povjerenstva)
• Committee member: Mihael Kovač (član povjerenstva)
• Committee member: Kristijan Poje (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2023-07-05, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Ovaj rad ispituje učinkovitost metode zvane tehnika lanca misli. Spomenuta tehnika temelji se na razvijanju strategije i višekoračnom postupku rješavanja problema, pri čemu se proces generiranja odgovora odvija pred korisnikom. Ispitivanje tehnike lanca misli provedeno je nad jednim od najpopularnijih velikih jezičnih modela – ChatGPT-jem. Rad se dotiče i obrađivanja prirodnog jezika te donosi pregled jezičnih modela i njihova razvoja. Istraživanje je provedeno pomoću naputaka koji opisuju scenarije s kakvima se susreću različite tvrtke. Primjenom tehnike lanca misli potaknuto je razvijanje poslovne strategije i njezin detaljan opis. U radu se evaluira generirani tekst te uspoređuje s očekivanim rješenjem. Uspoređuju se i rezultati GPT-3.5 i GPT-4 verzija. Konačno, predstavljene su metode koje unaprjeđuju rezultate generirane opisanom tehnikom.

Abstract (english)

This thesis examines the efficacy of the 'Chain-of-Thought' (CoT) technique, which is rooted in strategic development and a step-by-step problem-solving process, where solutions are generated in front of the user. The studied technique was evaluated using ChatGPT, one of the most popular large language models. The research further explores natural language processing, provides a short overview of language models and their evolution. By employing guidelines depicting various business scenarios, the technique stimulated the development of a business strategy. The paper assesses the generated text against the expected solution. Results from GPT-3.5 and GPT-4 versions are comparatively analysed. Finally, the thesis presents methods that enhance the results generated by this technique.

• Keywords: tehnika lanca misli; davanje naputaka; veliki jezični modeli; ChatGPT
• Keywords (english): Chain-of-Thought; prompting; Large Language Models; ChatGPT
• Language: croatian
• URN:NBN: urn:nbn:hr:168:207989
• Study programme: Title: Computing; Study programme type: university; Study level: undergraduate; Academic / professional title: sveučilišni/a prvostupnik/ prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka računarstva (sveučilišni/a prvostupnik/ prvostupnica (baccalaureus/baccalaurea) inženjer/inženjerka računarstva)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2023-07-13 10:10:03