Work

Adversarial Prompt Optimization in LLMs: HijackNet’s Approach to Robustness and Defense Evasion

Year: 2025

Type: conference-paper

Abstract: This paper introduces HijackNet, an adversarial prompt hijacking framework designed to improve hijacking success rates and generalization in multi-lingual tasks. Combining Neural Architecture Search (NAS), Reinforcement Learning (RL), and Multi-Objective Optimization (MOO), HijackNet generates adversarial prompts that adapt to target tasks while bypassing model defense mechanisms. This framework optimizes hijacking success, defense evasion, and task completion through a combination of a prompt generator and adversarial optimizer. We introduce a Defense Robustness Score (DRS) to evaluate the stability of prompts in adversarial settings. HijackNet improves adversarial prompt generation, advancing the security and robustness of large language models in adversarial environments.

DOI: 10.1109/iscait64916.2025.11010431

Authors: Lei Fu, Kowei Shih, Jinghan Cao, Siye Wu, Yihong Jin, Ze Yang

Institutions: Tsinghua University, San Francisco State University, University of Rochester, University of Illinois Urbana-Champaign

Language: English

Topic: Adversarial Robustness in Machine Learning

Subfield: Artificial Intelligence

Field: Computer Science

Domain: Physical Sciences

Open Access status: green

OpenAlex ID: W4410987182

FWCI:

Cites: 9

Cited by:

Related to: 10