Language Model Security Database

A vulnerability in several large language models (LLMs), including Qwen2.5-7BInstruct, Llama3.1-8B-Instruct, and GPT-4 variants, allows for black-box jailbreaking via prompt engineering techniques that exploit the proximity of benign and malicious prompt embeddings in the model's representation space. An attacker can craft prompts leveraging reinforcement learning to manipulate the embedding, causing the model to bypass its safety mechanisms and generate harmful or undesirable outputs while…

This vulnerability allows an attacker to bypass the safety mechanisms of Large Language Models (LLMs) by using an evolutionary algorithm to generate effective jailbreak prompts. The algorithm leverages the LLM's capabilities to iteratively refine prompts, increasing the likelihood of eliciting harmful responses to otherwise disallowed queries.

The GAP framework, as described in arXiv:2501.18638, reveals vulnerabilities in various large language models (LLMs) by generating stealthy jailbreak prompts that bypass content moderation systems. The framework leverages a graph-based attack strategy, enabling knowledge sharing across attack paths for enhanced efficiency and evasion. This allows the successful bypassing of multiple LLM safety mechanisms, including those based on perplexity and prompt-based heuristics.

The Virus attack method enables attackers to bypass guardrail moderation on fine-tuning data, leading to a significant degradation of safety alignment in large language models (LLMs). This is achieved through a dual-objective data optimization strategy that crafts harmful data undetectable by the guardrail while maximizing their effectiveness in compromising the victim model's safety.

Large language models (LLMs) exhibit increased responsiveness to prompts framed within positive narratives. The Happy Ending Attack (HEA) exploits this by embedding malicious requests within a positive-sentiment scenario culminating in a happy ending. This allows the LLM to generate responses that fulfill the malicious request while perceiving the overall prompt as benign.

End-to-end Large Audio Language Models (LALMs) contain an audio-based jailbreak vulnerability allowing attackers to bypass safety alignment guardrails by manipulating audio-specific "hidden semantics." Unlike text-based attacks, this exploitation involves encoding harmful queries into audio and applying signal processing modifications—specifically changes to emphasis, speech speed, intonation, tone, background noise, celebrity accents, or emotional overlays (e.g., laughter, screaming). These…

Large Language Models (LLMs) used in hate speech detection systems are vulnerable to adversarial attacks and model stealing, resulting in evasion of hate speech detection. Adversarial attacks modify hate speech text to evade detection, while model stealing creates surrogate models that mimic the target system's behavior.

Vision Language Models (VLMs) are vulnerable to visual prompt injection attacks via text-to-image obfuscation. While these models often possess safety guardrails for standard text-based inputs, they fail to apply equivalent safety alignment to textual instructions embedded visually within an image. An attacker can overlay malicious instructions (e.g., requests for illegal acts, hate speech) onto an image file and submit it to the model. The model’s Optical Character Recognition (OCR) or visual…

Large Language Model (LLM) based search engines utilizing Retrieval-Augmented Generation (RAG) are vulnerable to ranking manipulation attacks via indirect prompt injection. Adversaries can embed optimized adversarial triggers or crafted semantic patterns within external webpage content. When these manipulated documents are retrieved and integrated into the LLM's context window alongside a user query, the adversarial content disrupts the model's contextual understanding. This results in the LLM…

Voting-based Large Language Model (LLM) leaderboards, such as Chatbot Arena, are vulnerable to adversarial ranking manipulation due to insufficient response anonymity. While these systems obscure model identities during head-to-head comparisons to prevent bias, an attacker can de-anonymize the models with high accuracy (>95%) by analyzing response content. The attack functions in two stages: (1) Re-identification, where the attacker submits specific prompts (identity-probing or stylometric…