Copertina

FlexLLM: Exploring LLM Customization for Moving Target Defense on Black-Box LLMs Against Jailbreak Attacks

Salvato in:
Pubblicato in:arXiv.org (Dec 10, 2024), p. n/a
Autore principale: Chen, Bocheng
Altri autori: Guo, Hanqing, Qiben Yan
Pubblicazione:
Cornell University Library, arXiv.org
Soggetti:
Application programming interface
Prompt engineering
Black boxes
Large language models
Defense
Moving targets
Accesso online:Citation/Abstract
Full text outside of ProQuest
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!

MARC

LEADER 00000nab a2200000uu 4500
001 3143054177
003 UK-CbPIL
022 |a 2331-8422 
035 |a 3143054177 
045 0 |b d20241210 
100 1 |a Chen, Bocheng 
245 1 |a FlexLLM: Exploring LLM Customization for Moving Target Defense on Black-Box LLMs Against Jailbreak Attacks 
260 |b Cornell University Library, arXiv.org  |c Dec 10, 2024 
513 |a Working Paper 
520 3 |a Defense in large language models (LLMs) is crucial to counter the numerous attackers exploiting these systems to generate harmful content through manipulated prompts, known as jailbreak attacks. Although many defense strategies have been proposed, they often require access to the model's internal structure or need additional training, which is impractical for service providers using LLM APIs, such as OpenAI APIs or Claude APIs. In this paper, we propose a moving target defense approach that alters decoding hyperparameters to enhance model robustness against various jailbreak attacks. Our approach does not require access to the model's internal structure and incurs no additional training costs. The proposed defense includes two key components: (1) optimizing the decoding strategy by identifying and adjusting decoding hyperparameters that influence token generation probabilities, and (2) transforming the decoding hyperparameters and model system prompts into dynamic targets, which are continuously altered during each runtime. By continuously modifying decoding strategies and prompts, the defense effectively mitigates the existing attacks. Our results demonstrate that our defense is the most effective against jailbreak attacks in three of the models tested when using LLMs as black-box APIs. Moreover, our defense offers lower inference costs and maintains comparable response quality, making it a potential layer of protection when used alongside other defense methods. 
653 |a Application programming interface 
653 |a Prompt engineering 
653 |a Black boxes 
653 |a Large language models 
653 |a Defense 
653 |a Moving targets 
700 1 |a Guo, Hanqing 
700 1 |a Qiben Yan 
773 0 |t arXiv.org  |g (Dec 10, 2024), p. n/a 
786 0 |d ProQuest  |t Engineering Database 
856 4 1 |3 Citation/Abstract  |u https://www.proquest.com/docview/3143054177/abstract/embedded/ZKJTFFSVAI7CB62C?source=fedsrch 
856 4 0 |3 Full text outside of ProQuest  |u http://arxiv.org/abs/2412.07672