TombRaider: Entering the Vault of History to Jailbreak Large Language Models

TL;DR

TombRaider is a novel jailbreak method that exploits historical knowledge in large language models to bypass safety filters, revealing significant vulnerabilities and outperforming existing techniques.

Contribution

Introduces TombRaider, a new jailbreak approach using historical knowledge and dual agents, significantly improving attack success rates on LLMs.

Findings

Achieves nearly 100% success on bare models

Over 55.4% success against defenses

Outperforms state-of-the-art jailbreak techniques

Abstract

Warning: This paper contains content that may involve potentially harmful behaviours, discussed strictly for research purposes. Jailbreak attacks can hinder the safety of Large Language Model (LLM) applications, especially chatbots. Studying jailbreak techniques is an important AI red teaming task for improving the safety of these applications. In this paper, we introduce TombRaider, a novel jailbreak technique that exploits the ability to store, retrieve, and use historical knowledge of LLMs. TombRaider employs two agents, the inspector agent to extract relevant historical information and the attacker agent to generate adversarial prompts, enabling effective bypassing of safety filters. We intensively evaluated TombRaider on six popular models. Experimental results showed that TombRaider could outperform state-of-the-art jailbreak techniques, achieving nearly 100% attack success rates (ASRs) on bare models and maintaining over 55.4% ASR against defence mechanisms. Our findings highlight critical vulnerabilities in existing LLM safeguards, underscoring the need for more robust safety defences.