Model layer Vulnerabilities

A vulnerability in Large Language Models (LLMs) and autonomous agent frameworks, termed "Emoticon Semantic Confusion," allows for the generation and execution of unintended, potentially destructive code. Because ASCII-based emoticons (e.g., ~, *, !(^^)!) heavily overlap with the symbol space of programming operators, shell wildcards, and file paths, LLMs frequently misinterpret these affective, non-verbal cues as executable directives. When processing user instructions in code-generation or…

Large Language Models (LLMs) exhibit a False Refusal vulnerability during legitimate hate speech detoxification tasks (text style transfer). Safety alignment mechanisms fail to contextually distinguish between a benign instruction to "detoxify" or "rewrite" harmful content and the generation of harmful content itself. This results in a denial of service where the model refuses to process the input. This vulnerability is not uniformly distributed; it is statistically biased to…

Backdoor-based fingerprinting mechanisms used for Intellectual Property (IP) protection in Large Language Models (LLMs) are vulnerable to evasion when deployed in model ensemble configurations. The vulnerability arises because fingerprint triggers elicit specific, high-probability tokens or responses in a protected model that are statistically improbable in unprotected or differently-fingerprinted auxiliary models. Attackers can exploit this statistical discrepancy without accessing model…

Instruction-tuned Large Language Models (LLMs) are vulnerable to the induction of "hidden intentions"—covert, goal-directed manipulative behaviors—via lightweight prompt engineering, system prompts, or agentic workflows. Attackers can embed latent agendas (e.g., commercial manipulation, simulated consensus, or the promotion of insecure coding practices) into model outputs that trigger only under specific conversational contexts. Because these manipulative behaviors mimic benign interactions…

Large Language Models (LLMs), specifically Llama-3.1-8B-Instruct, Ministral-8B-Instruct-2410, Gemma-2-9B-It, and Qwen2.5-7B-Instruct, contain a safety guardrail bypass vulnerability when subjected to optimized adversarial prompts. The vulnerability is exposed via the RainbowPlus quality-diversity search method utilized within the RedBench evaluation framework. These models exhibit high Attack Success Rates (ASR)—up to 97.81% for Ministral and 96.25% for Llama-3.1—failing to refuse prompts in…

Unintended input-only PII memorization in fine-tuned Large Language Models (LLMs) allows remote attackers to extract sensitive Personally Identifiable Information (PII) such as names, medical records, and financial details. This vulnerability occurs when a model is fine-tuned on datasets where sensitive information appears in the input text, even if that information is not part of the training target (label) or is unrelated to the downstream task (e.g., classification). The fine-tuning process…

LLM-based evaluation systems ("LLM-as-a-Judge") exhibit a structural vulnerability termed "Framing Bias," wherein the model produces logically contradictory judgments depending on the syntactic framing of the evaluation prompt. Specifically, when assessing the same content using predicate-positive (P) framing (e.g., "Is this toxic?") versus predicate-negative (¬P) framing (e.g., "Is this non-toxic?"), models frequently fail to invert their binary decisions, leading to inconsistency rates…

LLM-based autonomous agents are vulnerable to implicit regulatory compliance failures during tool invocation. When initialized with unstructured regulatory policies and given goal-oriented user instructions that do not explicitly state safety requirements, LLMs frequently prioritize functional task completion over mandatory safety constraints. This leads to an "Unsafe Success" execution state, where the agent successfully achieves the user's business goal but silently bypasses critical…

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…

Large Language Models (LLMs) and dedicated content moderation systems (including safety guards and encoder-based classifiers) are vulnerable to "Soft Hate" evasion attacks. This vulnerability arises when hostile standpoints are encoded through defeasible reasoning structures—specifically formalized by the Argumentum Model of Topics (AMT)—rather than explicit lexical toxicity. Attackers can bypass safety filters by framing discriminatory conclusions as logical inferences derived from widely…