Black-box Vulnerabilities

Jailbreaking vulnerabilities in Large Language Models (LLMs) used in Retrieval-Augmented Generation (RAG) systems allow escalation of attacks from entity extraction to full document extraction and enable the propagation of self-replicating malicious prompts ("worms") within interconnected RAG applications. Exploitation leverages prompt injection to force the LLM to return retrieved documents or execute malicious actions specified within the prompt.

PathSeeker demonstrates a novel black-box jailbreak attack against Large Language Models (LLMs) that utilizes multi-agent reinforcement learning. The attack iteratively modifies input prompts based on model responses, leveraging a reward mechanism focused on vocabulary expansion in the LLM's output to circumvent safety mechanisms and elicit harmful responses. This technique bypasses existing safety filters by encouraging the model to relax its constraints, rather than directly targeting…

A single-turn prompt injection attack that bypasses LLM content moderation filters by simulating a multi-turn conversation escalating towards harmful or inappropriate outputs within a single prompt. The attack leverages the LLM's tendency to maintain context and continue established patterns, even when leading to undesirable content.

Large Language Models (LLMs) are vulnerable to a jailbreaking attack, termed "MathPrompt," which leverages the models' ability to process symbolic mathematics to bypass built-in safety mechanisms. The attack encodes harmful natural language prompts into mathematically formulated problems, causing the LLM to generate unsafe outputs while ostensibly solving a mathematical problem.

Large Language Models (LLMs) are vulnerable to a targeted jailbreak attack, termed Atoxia, which can force the generation of specific harmful content. The attack operates by providing a target toxic answer to an attacker model, which then generates a corresponding adversarial query and a misleading "answer opening" (prefix). When the query and the answer prefix are presented to a vulnerable LLM, the model is induced to continue the generation, bypassing its safety alignment and completing the…

Large Language Models (LLMs) are vulnerable to a novel jailbreak attack that leverages "neural carrier articles." This attack injects a prohibited query into a benign article generated by a secondary LLM, designed to be semantically similar to the prohibited query but not trigger the primary LLM's safety mechanisms. The secondary LLM generates articles based on hypernyms derived from the prohibited query, thus subtly shifting attention weights within the primary LLM, bypassing its safeguards.

Large Language Models (LLMs) are vulnerable to composable jailbreak attacks, allowing bypass of safety filters through the chaining of multiple prompt transformations. The vulnerability arises from the ability to combine seemingly innocuous transformations to create effective attacks that achieve high attack success rates (ASR). These attacks can be synthesized automatically, allowing for the creation of novel and highly effective jailbreaks. Specifically, using the h4rm3l framework, attacks…

Large Language Models (LLMs) are vulnerable to a multi-turn context-based jailbreak attack, termed Context Fusion Attack (CFA). CFA leverages the LLM's ability to understand context in multi-turn dialogues to bypass security mechanisms designed to prevent harmful outputs. The attack involves strategically crafting a series of prompts that build context, subtly introducing malicious keywords, and ultimately triggering the LLM to generate unsafe content. The malicious intent is masked within the…

The Ensemble Jailbreak (EnJa) attack exploits vulnerabilities in the safety mechanisms of large language models (LLMs) by combining prompt-level and token-level attacks. EnJa conceals malicious instructions within seemingly benign prompts, then uses a gradient-based method to optimize adversarial suffixes, significantly increasing the likelihood of bypassing safety filters and generating harmful content. The attack leverages a connector template to seamlessly integrate the concealed prompt and…

A Cross-Prompt Injection Attack (XPIA) can be amplified by appending a Greedy Coordinate Gradient (GCG) suffix to the malicious injection. This increases the likelihood that a Large Language Model (LLM) will execute the injected instruction, even in the presence of a user's primary instruction, leading to data exfiltration. The success rate of the attack depends on the LLM's complexity; medium-complexity models show increased vulnerability.