Original title: Nová obranná metoda proti jailbreak útokům na velké jazykové modely
Translated title: A Novel Defense Method against Jailbreak Attacks on Large Language Models
Authors: Kaška, Petr ; Firc, Anton (referee) ; Reš, Jakub (advisor)
Document type: Master’s theses
Year: 2026
Language: eng
Publisher: Vysoké učení technické v Brně. Fakulta informačních technologií
Abstract: [eng] [cze]

Keywords: adversariální prompty; bezpečnost umělé inteligence; genetické programování; jailbreak útoky; LLM jako hodnotitel; prompt-level obrana; velké jazykové modely; znaková perturbace; adversarial prompts; AI safety; character perturbation; genetic programming; jailbreak attacks; large language models; LLM-as-judge; prompt-level defense

Institution: Brno University of Technology (web)
Document availability information: Fulltext is available in the Brno University of Technology Digital Library.
Original record: http://hdl.handle.net/11012/260425

Permalink: http://www.nusl.cz/ntk/nusl-776267


The record appears in these collections:
Universities and colleges > Public universities > Brno University of Technology
Academic theses (ETDs) > Master’s theses
 Record created 2026-06-27, last modified 2026-07-24


No fulltext
  • Export as DC, NUŠL, RIS
  • Share