Language Model Security Database

Generative reward models deployed as LLM-as-a-Judge (LaaJ) evaluators contain a logic bypass vulnerability where superficial "master key" inputs trigger false positive rewards regardless of actual response quality. Instead of evaluating the candidate's output, large judge models are inadvertently triggered by specific token sequences to solve the prompt independently. This allows malicious actors or policy models undergoing reinforcement learning to consistently game the reward signal by…

Hybrid monitoring protocols for Large Language Model (LLM) agents, specifically those that strictly evaluate both Chain-of-Thought (CoT) reasoning and tool usage (such as LlamaFirewall's AlignmentCheck and Extract-and-Evaluate monitors), are vulnerable to "Agent-as-a-Proxy" indirect prompt injection attacks. Unlike standard injections that target the agent, this attack targets the monitoring model itself. By embedding a malicious string—optimized via an algorithm named Parallel-GCG—into…

Large Language Models (LLMs) integrated with external retrieval mechanisms (e.g., Retrieval-Augmented Generation (RAG), web search, or email processing) are vulnerable to Indirect Prompt Injection. This vulnerability occurs when an LLM consumes input from untrusted external sources—such as websites, code repositories, or incoming emails—that contain embedded adversarial prompts. Unlike direct injection, where the user attacks the model, here the "poisoned" data is retrieved by the system…

The GPT-OSS-20B large language model contains critical failures in its alignment and Chain-of-Thought (CoT) reasoning mechanisms, specifically in how it prioritizes numerical objectives and validates procedural structure. The model is vulnerable to "Quant Fever," where explicit numerical targets in a prompt (e.g., "delete 90% of files") override contextual safety constraints (e.g., "do not delete important files"). Furthermore, the model exhibits "Reasoning Procedure Mirage," where harmful…

Adversarial Activation Patching enables the induction of emergent deceptive behaviors in safety-aligned transformer-based Large Language Models (LLMs). By extracting intermediate activations ($A_{d}$) generated during the processing of a deceptive or harmful prompt and injecting them into the forward pass of a benign target prompt ($x_{t}$) at specific layers (specifically mid-layers, e.g., 5-10 in 32-layer architectures), an attacker can manipulate the model's internal reasoning circuits…

Large Language Models (LLMs) utilizing Chain-of-Thought (CoT) prompting are vulnerable to input perturbations that decouple intermediate reasoning from the final answer. An attacker can generate adversarial examples using gradient-based optimization (targeting specific loss functions that maximize reasoning divergence while minimizing answer loss) to induce "Right Answer, Wrong Reasoning" behaviors. This vulnerability manifests through two primary attack vectors: 1. Token-level perturbations…

Predictive Large Language Model (LLM) routers, specifically those utilizing Deep Neural Network (DNN) and Matrix Factorization (MF) architectures, are vulnerable to adversarial manipulation and backdoor poisoning. These routers are designed to optimize cost and latency by dynamically directing simple queries to "weak" (cheap) models and complex queries to "strong" (expensive) models. Attackers can exploit this mechanism in two ways: 1. Inference-time Attacks: By appending specific adversarial…

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.