| Naslov: | Detection of malicious software using large language models : master's degree thesis |
|---|
| Avtorji: | ID Tivadar, Martina (Avtor)
ID Karakatič, Sašo (Mentor) Več o mentorju...  |
|---|
| Datoteke: |  MAG_Tivadar_Martina_2025.pdf (29,72 MB)
MD5: 9D547BEA2188C05F23797101E8DADDD1
|
|---|
| Jezik: | Angleški jezik |
|---|
| Vrsta gradiva: | Magistrsko delo/naloga |
|---|
| Tipologija: | 2.09 - Magistrsko delo |
|---|
| Organizacija: | FERI - Fakulteta za elektrotehniko, računalništvo in informatiko
|
|---|
| Opis: | This thesis examines the success rate of large language models (LLM) in detecting macOS malware through Endpoint Security logs. A literature review and 144 experiments with three ChatGPT variants and six prompt types evaluated accuracy, precision, recall, specificity, and F1-score. Results show that prompt wording is crucial: zero-shot and chain-of-thought prompts performed best, while conservative prompts minimized false positives but missed threats. GPT-4o and o1 outperformed o4-mini but showed similar results. Findings suggest LLMs can support, but not replace, traditional detection, with prompt design proving as important as model choice. |
|---|
| Ključne besede: | malware, large language models, detection |
|---|
| Kraj izida: | Maribor |
|---|
| Kraj izvedbe: | Maribor |
|---|
| Založnik: | [M. Tivadar] |
|---|
| Leto izida: | 2025 |
|---|
| Št. strani: | 1 spletni vir (1 datoteka PDF (XVI, 76 str.)) |
|---|
| PID: | 20.500.12556/DKUM-95400 |
|---|
| UDK: | 004.056.54(043.2) |
|---|
| COBISS.SI-ID: | 262071811 |
|---|
| Datum objave v DKUM: | 03.11.2025 |
|---|
| Število ogledov: | 290 |
|---|
| Število prenosov: | 36 |
|---|
| Metapodatki: |    |
|---|
| Področja: | KTFMB - FERI
|
|---|
|
:
|
Kopiraj citat |
|---|
| | | | Skupna ocena: | (0 glasov) |
|---|
| Vaša ocena: | Ocenjevanje je dovoljeno samo prijavljenim uporabnikom. |
|---|
| Objavi na: |  |
|---|
Postavite miškin kazalec na naslov za izpis povzetka. Klik na naslov izpiše
podrobnosti ali sproži prenos. |
