LLMs Caught in the Crossfire: Malware Requests and Jailbreak Challenges

TL;DR

This paper introduces MalwareBench, a benchmark dataset to evaluate LLMs' vulnerability to jailbreak attacks in malicious code generation, revealing significant security challenges in current models.

Contribution

We created MalwareBench, a comprehensive benchmark with 3,520 prompts covering multiple jailbreak methods, to systematically assess LLM security against malicious code generation.

Findings

Mainstream LLMs have limited ability to reject malicious code requests.

Combining multiple jailbreak methods significantly reduces model security.

Average rejection rate drops to 39.92% with combined jailbreak attacks.

Abstract

The widespread adoption of Large Language Models (LLMs) has heightened concerns about their security, particularly their vulnerability to jailbreak attacks that leverage crafted prompts to generate malicious outputs. While prior research has been conducted on general security capabilities of LLMs, their specific susceptibility to jailbreak attacks in code generation remains largely unexplored. To fill this gap, we propose MalwareBench, a benchmark dataset containing 3,520 jailbreaking prompts for malicious code-generation, designed to evaluate LLM robustness against such threats. MalwareBench is based on 320 manually crafted malicious code generation requirements, covering 11 jailbreak methods and 29 code functionality categories. Experiments show that mainstream LLMs exhibit limited ability to reject malicious code-generation requirements, and the combination of multiple jailbreak methods further reduces the model's security capabilities: specifically, the average rejection rate for malicious content is 60.93%, dropping to 39.92% when combined with jailbreak attack algorithms. Our work highlights that the code security capabilities of LLMs still pose significant challenges.