EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Evaluating Prompt Injection Attacks with LSTM-Based Generative Adversarial Networks: A Lightweight Alternative to Large Language Models

Sharaf Rashid, Edson Bollis, Lucas Pellicer, Darian Rabbani, Rafael Palacios, Aneesh Gupta, Amar Gupta

Machine Learning and Knowledge Extraction · 06.08.2025 · 7 (3) · 77

DOI: 10.3390/make7030077 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Machine Learning and Knowledge Extraction, online erschienen am 06.08.2025. Verfasst von Sharaf Rashid und 6 weiteren Autor:innen. Die Arbeit wurde bisher 6-mal zitiert (Crossref) und nennt 52 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Instituto de Ciência e Tecnologia Itaú (ICTi).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Generative Adversarial Networks (GANs) using Long Short-Term Memory (LSTM) provide a computationally cheaper approach for text generation compared to large language models (LLMs). The low hardware barrier of training GANs poses a threat because it means more bad actors may use them to mass-produce prompt attack messages against LLM systems. Thus, to better understand the threat of GANs being used for prompt attack generation, we train two well-known GAN architectures, SeqGAN and RelGAN, on prompt attack messages. For each architecture, we evaluate generated prompt attack messages, comparing results with each other, with generated attacks from another computationally cheap approach, a 1-billion-parameter Llama 3.2 small language model (SLM), and with messages from the original dataset. This evaluation suggests that GAN architectures like SeqGAN and RelGAN have the potential to be used in conjunction with SLMs to readily generate malicious prompts that impose new threats against LLM-based systems such as chatbots. Analyzing the effectiveness of state-of-the-art defenses against prompt attacks, we also find that GAN-generated attacks can deceive most of these defenses with varying levels of success with the exception of Meta’s PromptGuard. Further, we suggest an improvement of prompt attack defenses based on the analysis of the language quality of the prompts, which we found to be the weakest point of GAN-generated messages.

Förderung

  • Instituto de Ciência e Tecnologia Itaú (ICTi)

Zitieren

APA

Rashid, S., Bollis, E., Pellicer, L., Rabbani, D., Palacios, R., Gupta, A., & Gupta, A. (2025). Evaluating Prompt Injection Attacks with LSTM-Based Generative Adversarial Networks: A Lightweight Alternative to Large Language Models. Machine Learning and Knowledge Extraction, 7(3), 77. https://doi.org/10.3390/make7030077

Vancouver

Rashid S, Bollis E, Pellicer L, Rabbani D, Palacios R, Gupta A, et al. Evaluating Prompt Injection Attacks with LSTM-Based Generative Adversarial Networks: A Lightweight Alternative to Large Language Models. Machine Learning and Knowledge Extraction. 2025;7(3):77. doi:10.3390/make7030077

BibTeX
@article{Rashid2025,
  author = {Rashid, Sharaf and Bollis, Edson and Pellicer, Lucas and Rabbani, Darian and Palacios, Rafael and Gupta, Aneesh and Gupta, Amar},
  title = {Evaluating Prompt Injection Attacks with LSTM-Based Generative Adversarial Networks: A Lightweight Alternative to Large Language Models},
  journal = {Machine Learning and Knowledge Extraction},
  year = {2025},
  volume = {7},
  number = {3},
  pages = {77},
  doi = {10.3390/make7030077}
}
Kostenlos starten →