Breaching Large Language Models

Research areas:

• Volume 13, Issue 4, Jan. 2025

Year:

2025

Type of Publication:

Article

Keywords:

Breaching Large Language Models, Large Language Models

Authors:

Dhruv Malik; Pratik Agarwal; Stefan Schmidt; Sanjay Patel

Journal:

IJAIM

Volume:

13

Number:

4

Pages:

7-14

Month:

January

ISSN:

2320-5121

Abstract:

Large language models (LLMs) have demonstrated unprecedented capabilities in natural language understanding and generation. However, these models are not impervious to adversarial attacks, which can expose and exploit their vulnerabilities. This paper presents a comprehensive survey of the various adversarial attacks targeting LLMs, categorizing them based on their techniques and objectives. We also highlight existing defense mechanisms and their effectiveness. Through this survey, we aim to provide a clearer understanding of the current landscape and offer insights into areas needing further research and development

Full text: IJAIM_679_FINAL.pdf