Aplikacija za inženjering upita razgovornih agenata

undergraduate thesis

• Title (english): Application for chatbot prompt engineering
• Author: Josip Šare
• Mentor: Marko Horvat (mentor)
• Committee member: Marko Horvat (predsjednik povjerenstva)
• Committee member: Kristina Ćaleta (član povjerenstva)
• Committee member: Andrej Slapničar (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2024-07-03, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing

Abstract

Inženjering upita je postupak ugrađivanja opisa željenog zadatka u dijelove upit koji se šalje razgovornom agentu odnosno LLM-u. Postoje četiri glavna dijela upita a to su instrukcija, kontekst, ulazni podatci i indikator izlaznih podataka. Pisanje dobrog upita svodi se na sažimanje što više informacija u upit za što je potrebno poznavanje područja znatiželje.
Implementacija ove aplikacije koristi dva LLM API-a od dvije od najistaknutijih kompanija u prostoru AI, OpenAI i Meta. Aplikacija omogućuje korisniku izbor korištenja jednog od dva ponuđena LLM modela, ChatGPT 3.5-Turbo u LLama-13b-chat. Korisnik upisuje svoj željeni upit te odabirom opcija specificira detalje finalnog upita. Korisnik na posljetku dobiva unaprijeđeni upit od programa te odgovor na taj isti upit.

Abstract (english)

Prompt engineering is the process of embedding the description of the desired task into parts of the prompt sent to the conversational agent or LLM. There are four main parts of a prompt: instruction, context, input data, and output indicator. Writing a good prompt involves condensing as much information as possible into the prompt, which requires knowledge of the area of interest.
The implementation of this application uses two LLM APIs from two of the most prominent companies in the AI space, OpenAI and Meta. The application allows the user to choose between two LLM models, ChatGPT 3.5-Turbo and LLama-13b-chat. The user enters their desired prompt and specifies the details of the final query by selecting options. Finally, the user receives an enhanced query from the program and the response to that same prompt.

• Keywords: Inženjering upita; projektiranje upita; vanjsko sučelje LLM API; nadopunjeni upit; Java Spring Boot
• Keywords (english): Prompt engineering; prompt design; external interface LLM API; enhanced query; Java Spring Boot
• Language: croatian
• URN:NBN: urn:nbn:hr:168:894671
• 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: Open access
• Created on: 2024-10-12 22:12:43