Language Model Security Database

A vulnerability exists in Large Language Model (LLM) and Large Reasoning Model (LRM) serving interfaces that allow user-defined response prefixes, such as plain text-completion (v1/completions), Fill-in-the-Middle (FIM), or assistant message prefilling. An attacker can perform a Response Prefix Attack (RPA) by injecting maliciously crafted Chain-of-Thought (CoT) reasoning tokens immediately following the assistant's start delimiter (e.g., <|im_start|>assistant). Because these tokens are placed…

Reasoning-specialized Large Language Models (LLMs) that utilize Chain-of-Thought (CoT) processes are vulnerable to reasoning-exploitation jailbreaks. Attackers can bypass standard safety alignments (such as RLHF) by using adaptive multi-turn interactions or semantic transformations to induce the model to generate intermediate reasoning steps that "rationalize" or "contextualize" a harmful request. Because current alignment techniques often fail to scale linearly with reasoning depth, forcing…

Large Language Models (LLMs), specifically variants of GPT-4o, DeepSeek-R1, OLMo-2, and Llama-4, are vulnerable to accelerated adaptive adversarial attacks due to excessive information leakage in observable output signals. When these models expose "thinking processes" (Chain-of-Thought traces) or token-level log-probabilities (logits) to the end user, they leak significant mutual information $I(Z;T)$ regarding the model's safety state or hidden instructions. This leakage allows adaptive attack…

Large Language Models (LLMs) aligned via techniques such as Reinforcement Learning with Human Feedback (RLHF) or Direct Preference Optimization (DPO) contain a vulnerability in how safety features are encoded within the model parameters. The safety-critical information is primarily stored in low-rank subspaces of the weight matrices (specifically, the difference between the base and aligned model weights). These low-rank subspaces are highly sensitive to parameter updates. Consequently…

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.

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.

Large Language Models (LLMs) are vulnerable to a novel attack leveraging subconscious exploitation and echopraxia. Attackers craft prompts that subtly guide the LLM to echo malicious content it has implicitly learned during pre-training but is programmed to suppress. This bypasses safety mechanisms designed to prevent the generation of harmful content. The technique involves extracting malicious knowledge from the LLM's conditional probability distribution (representing its "subconscious") and…