Skip to main content
LLM Security Database
Skip to research search
Last analyzed Sep 9, 2026

Language Model Security Database

985 research findings · 1119 evaluated models

Filtered research findings

468 entries

Matches every word across titles, descriptions, sources, affected systems, and models.

Strategic Abstention Failure

Published Analyzed Source: arXiv

A "risk-invariance" vulnerability exists in Large Language Models (LLMs) wherein the model's decision-making policy is functionally decoupled from its verbalized confidence and externally defined error penalties. Despite generating calibrated confidence estimates (internal epistemic uncertainty), affected models fail to adjust their abstention thresholds when presented with high-penalty incentives (e.g., negative utility for incorrect answers). This results in "utility collapse," where models…

Are LLM Decisions Faithful to Verbal Confidence?
Evaluated models: GPT-5 Mini, GPT-5 Nano, GPT-4.1 Mini+7 more
  • GPT-5 Mini
  • GPT-5 Nano
  • GPT-4.1 Mini
  • Gemini 3 Flash
  • Gemini 2.5 Flash
  • Llama 4 Maverick
  • Gemma 3n E4B
  • DeepSeek V3.2
  • DeepSeek V3.2 Thinking
  • Qwen 3 Next 80B-A3B Thinking

Adversarial Prompts Defeat Code Defenses

Published Analyzed Source: arXiv

State-of-the-art secure code generation methods (Sven, SafeCoder, and PromSec) are vulnerable to adversarial prompt perturbations during inference, allowing for the bypass of security alignment mechanisms. The vulnerability stems from the models' reliance on surface-level textual pattern matching rather than semantic security reasoning. By employing simple prompt manipulations—such as Cue Inversion (flipping security directives), Naturalness Reframing (rewriting comments as novice questions)…

How Secure is Secure Code Generation? Adversarial Prompts Put LLM Defenses to the Test
Evaluated models: GPT-3.5, GPT-4o, Mistral 7B

LLM Review Paraphrase Attack

Published Analyzed Source: arXiv

LLM-as-a-Reviewer systems, which utilize large language models to automate the peer review process, are vulnerable to the Paraphrasing Adversarial Attack (PAA). PAA is a black-box optimization technique that exploits the model's sensitivity to specific input sequences and self-preference bias. By iteratively paraphrasing specific manuscript sections (such as the abstract) using in-context learning (ICL) guided by previous review scores, an attacker can generate adversarial sequences that…

Paraphrasing Adversarial Attack on LLM-as-a-Reviewer
Evaluated models: GPT-4o, Claude Sonnet 4

Point Cloud Adversarial Attack

Published Analyzed Source: arXiv

Point-based 3D Vision-Language Models (VLMs), specifically PointLLM and GPT4Point, are vulnerable to white-box, gradient-based adversarial attacks. The vulnerability exists in the model's processing of 3D point cloud data, where an attacker can optimize imperceptible geometric perturbations ($\delta$) on the input point cloud ($x$) to manipulate the model's textual output. The paper identifies two specific attack vectors: 1. Vision Attack: Directly perturbs the high-dimensional visual token…

On the Adversarial Robustness of 3D Large Vision-Language Models
Evaluated models: Vicuna 7B

Sophisticated Deception Induces Misbelief

Published Analyzed Source: arXiv

Large Language Models (LLMs) exhibit a vulnerability to "hard-to-falsify" deceptive evidence injection, termed the "Facade of Truth." This vulnerability allows an attacker to override an LLM’s parametric knowledge (internal factual beliefs) by injecting sophisticated, iteratively refined fabricated evidence into the context window. Unlike overt misinformation which models typically reject, this attack utilizes a multi-agent adversarial framework (MisBelief) to generate evidence that mimics…

The Facade of Truth: Uncovering and Mitigating LLM Susceptibility to Deceptive Evidence
Evaluated models: GPT-3.5, GPT-5, Llama 3 8B+1 more
  • GPT-3.5
  • GPT-5
  • Llama 3 8B
  • Qwen 2.5 32B

FusionRoute Logit Poisoning

Published Analyzed Source: arXiv

A fundamental algorithmic limitation exists in purely token-level multi-LLM collaboration systems (such as the "Collab" framework or routing-only variants of FusionRoute) that rely solely on selecting fixed expert outputs without complementary generation. The vulnerability, formally defined as an Identifiability Failure in Token-Level Routing, arises because observing optimal state-action values ($Q^$) along trajectories is insufficient to uniquely identify the specific expert action required…

Token-Level LLM Collaboration via FusionRoute
Evaluated models: GPT-4o, Llama 3.1 8B, Mistral 7B+1 more
  • GPT-4o
  • Llama 3.1 8B
  • Mistral 7B
  • Gemma 2 2B

Knowledge-Graph Implicit Prompts

Published Analyzed Source: arXiv

Large Language Models (LLMs) are vulnerable to a domain-specific obfuscation attack method termed "StealthGraph," which leverages Knowledge Graph (KG) guidance to bypass safety alignment. The vulnerability arises because current safety mechanisms primarily focus on explicit, general-domain harmful queries and fail to generalize to implicit, highly technical requests in specialized domains (e.g., medicine, finance, law).

StealthGraph: Exposing Domain-Specific Risks in LLMs through Knowledge-Graph-Guided Harmful Prompt Generation
Evaluated models: GPT-4o Mini, Gemini 2.5 Flash, Grok 3 Mini+10 more
  • GPT-4o Mini
  • Gemini 2.5 Flash
  • Grok 3 Mini
  • DeepSeek V3.1
  • Mixtral 8x7B
  • Qwen 2.5 7B
  • Llama 3.1 8B
  • Llama 3.1 70B
  • Qwen 3.5 Plus
  • GLM-5
  • Gemini 3 Pro
  • Llama Guard 4 12B
  • Vicuna 13B v1.5

LLM Conspiracy Bunking

Published Analyzed Source: arXiv

OpenAI GPT-4o is vulnerable to a targeted persuasion attack where the model acts as an active advocate for conspiracy theories. Standard safety guardrails do not prevent the model from generating specious, invented, or misleading arguments to successfully increase user belief in false claims (a "bunking" attack). Additionally, when explicitly constrained by system prompts to use only truthful information, the model adapts by "paltering"—strategically omitting context, juxtaposing true claims…

Large language models can effectively convince people to believe conspiracies
Evaluated models: GPT-4, GPT-4o

LLM Watermark Translation Bypass

Published Analyzed Source: arXiv

Token-level embedding-time watermarking algorithms, specifically KGW (Kirchenbauer et al.) and Exponential Sampling (EXP, Kuditipudi et al.), when implemented in Large Language Models (LLMs) for Bangla text generation, are vulnerable to watermark erasure via cross-lingual round-trip translation (RTT) attacks. While these methods achieve high detection accuracy (>88%) under benign conditions, translating watermarked Bangla text to English and back to Bangla causes detection accuracy to collapse…

BanglaLorica: Design and Evaluation of a Robust Watermarking Algorithm for Large Language Models in Bangla Text Generation
Evaluated models: Llama 3 8B

Multi-turn MLLM Jailbreak

Published Analyzed Source: arXiv

Multi-modal Large Language Models (MLLMs) are vulnerable to a multi-turn jailbreaking attack that leverages typographic visual prompts combined with conversational context drifting. The vulnerability exists because MLLMs establish trust and context during initial benign interactions, shifting the model's latent representation toward helpfulness and compromising its ability to detect malicious intent in subsequent turns. The attack vector utilizes an image where a harmful request is…

Multi-turn Jailbreaking Attack in Multi-Modal Large Language Models
Evaluated models: GPT-4o, Gemini 2.0 Flash, Qwen2-VL 7B Instruct+2 more
  • GPT-4o
  • Gemini 2.0 Flash
  • Qwen2-VL 7B Instruct
  • LLaVA 1.6 Mistral 7B
  • LLaVA 1.5 13B

Research methodology

Entries summarize publicly available primary-source security research. Model names reflect only systems explicitly evaluated by the cited paper, and measurements are research-reported unless independent verification is stated.

Discovery and publication are separate: new papers enter a candidate queue, then require human review and explicit approval before publication. Updates follow that review process, not a guaranteed daily publishing schedule. “Last analyzed” is the newest analysis date among published entries; it does not mean that all recent research has been covered.