InfoFlood: Jailbreaking Large Language Models with Information Overload

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Yadav, Advait, Jin, Haibo, Luo, Man, Zhuang, Jun, Wang, Haohan
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911006452613120
author Yadav, Advait
Jin, Haibo
Luo, Man
Zhuang, Jun
Wang, Haohan
author_facet Yadav, Advait
Jin, Haibo
Luo, Man
Zhuang, Jun
Wang, Haohan
contents Large Language Models (LLMs) have demonstrated remarkable capabilities across various domains. However, their potential to generate harmful responses has raised significant societal and regulatory concerns, especially when manipulated by adversarial techniques known as "jailbreak" attacks. Existing jailbreak methods typically involve appending carefully crafted prefixes or suffixes to malicious prompts in order to bypass the built-in safety mechanisms of these models. In this work, we identify a new vulnerability in which excessive linguistic complexity can disrupt built-in safety mechanisms-without the need for any added prefixes or suffixes-allowing attackers to elicit harmful outputs directly. We refer to this phenomenon as Information Overload. To automatically exploit this vulnerability, we propose InfoFlood, a jailbreak attack that transforms malicious queries into complex, information-overloaded queries capable of bypassing built-in safety mechanisms. Specifically, InfoFlood: (1) uses linguistic transformations to rephrase malicious queries, (2) identifies the root cause of failure when an attempt is unsuccessful, and (3) refines the prompt's linguistic structure to address the failure while preserving its malicious intent. We empirically validate the effectiveness of InfoFlood on four widely used LLMs-GPT-4o, GPT-3.5-turbo, Gemini 2.0, and LLaMA 3.1-by measuring their jailbreak success rates. InfoFlood consistently outperforms baseline attacks, achieving up to 3 times higher success rates across multiple jailbreak benchmarks. Furthermore, we demonstrate that commonly adopted post-processing defenses, including OpenAI's Moderation API, Perspective API, and SmoothLLM, fail to mitigate these attacks. This highlights a critical weakness in traditional AI safety guardrails when confronted with information overload-based jailbreaks.
format Preprint
id arxiv_https___arxiv_org_abs_2506_12274
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle InfoFlood: Jailbreaking Large Language Models with Information Overload
Yadav, Advait
Jin, Haibo
Luo, Man
Zhuang, Jun
Wang, Haohan
Cryptography and Security
Computation and Language
Large Language Models (LLMs) have demonstrated remarkable capabilities across various domains. However, their potential to generate harmful responses has raised significant societal and regulatory concerns, especially when manipulated by adversarial techniques known as "jailbreak" attacks. Existing jailbreak methods typically involve appending carefully crafted prefixes or suffixes to malicious prompts in order to bypass the built-in safety mechanisms of these models. In this work, we identify a new vulnerability in which excessive linguistic complexity can disrupt built-in safety mechanisms-without the need for any added prefixes or suffixes-allowing attackers to elicit harmful outputs directly. We refer to this phenomenon as Information Overload. To automatically exploit this vulnerability, we propose InfoFlood, a jailbreak attack that transforms malicious queries into complex, information-overloaded queries capable of bypassing built-in safety mechanisms. Specifically, InfoFlood: (1) uses linguistic transformations to rephrase malicious queries, (2) identifies the root cause of failure when an attempt is unsuccessful, and (3) refines the prompt's linguistic structure to address the failure while preserving its malicious intent. We empirically validate the effectiveness of InfoFlood on four widely used LLMs-GPT-4o, GPT-3.5-turbo, Gemini 2.0, and LLaMA 3.1-by measuring their jailbreak success rates. InfoFlood consistently outperforms baseline attacks, achieving up to 3 times higher success rates across multiple jailbreak benchmarks. Furthermore, we demonstrate that commonly adopted post-processing defenses, including OpenAI's Moderation API, Perspective API, and SmoothLLM, fail to mitigate these attacks. This highlights a critical weakness in traditional AI safety guardrails when confronted with information overload-based jailbreaks.
title InfoFlood: Jailbreaking Large Language Models with Information Overload
topic Cryptography and Security
Computation and Language
url https://arxiv.org/abs/2506.12274