Feint and Attack: Attention-Based Strategies for Jailbreaking and Protecting LLMs

TL;DR

This paper introduces attention-based metrics and strategies to analyze, attack, and defend large language models by manipulating their attention distributions, revealing vulnerabilities and proposing robustness enhancements.

Contribution

It proposes novel attention metrics and the ABA/ABD strategies, combining analysis, attack, and defense methods to improve understanding and security of LLMs.

Findings

ABA effectively diverts attention to induce harmful outputs.

ABD enhances LLM robustness by calibrating attention distribution.

Attention distribution significantly impacts LLM output quality.

Abstract

Jailbreak attack can be used to access the vulnerabilities of Large Language Models (LLMs) by inducing LLMs to generate the harmful content. And the most common method of the attack is to construct semantically ambiguous prompts to confuse and mislead the LLMs. To access the security and reveal the intrinsic relation between the input prompt and the output for LLMs, the distribution of attention weight is introduced to analyze the underlying reasons. By using statistical analysis methods, some novel metrics are defined to better describe the distribution of attention weight, such as the Attention Intensity on Sensitive Words (Attn_SensWords), the Attention-based Contextual Dependency Score (Attn_DepScore) and Attention Dispersion Entropy (Attn_Entropy). By leveraging the distinct characteristics of these metrics, the beam search algorithm and inspired by the military strategy "Feint and Attack", an effective jailbreak attack strategy named as Attention-Based Attack (ABA) is proposed. In the ABA, nested attack prompts are employed to divert the attention distribution of the LLMs. In this manner, more harmless parts of the input can be used to attract the attention of the LLMs. In addition, motivated by ABA, an effective defense strategy called as Attention-Based Defense (ABD) is also put forward. Compared with ABA, the ABD can be used to enhance the robustness of LLMs by calibrating the attention distribution of the input prompt. Some comparative experiments have been given to demonstrate the effectiveness of ABA and ABD. Therefore, both ABA and ABD can be used to access the security of the LLMs. The comparative experiment results also give a logical explanation that the distribution of attention weight can bring great influence on the output for LLMs.