Safety Vulnerabilities

Diffusion Large Language Models (D-LLMs) are vulnerable to a "Context Nesting" attack that bypasses safety alignment mechanisms. While D-LLMs typically utilize a stepwise reduction effect during the iterative denoising process to suppress harmful content, this mechanism fails when harmful requests are embedded within benign, structured contexts. By wrapping a malicious query inside high-level structural templates (such as code completion, table filling, JSON, or YAML formats), an attacker can…

Discrete image tokenizers are vulnerable to unsupervised embedding-space adversarial attacks. Attackers can apply $\ell_p$-bounded perturbations to an input image to maximize the $\ell_2$ distance of the pre-quantization continuous embeddings produced by the tokenizer's vision encoder. This forces the vector quantizer to cross discrete cell boundaries and assign incorrect codebook vectors, fundamentally altering the resulting token sequence. Because the attack targets the pre-quantization…

Large Language Models (LLMs) aligned via standard Reinforcement Learning from Human Feedback (RLHF) or Supervised Fine-Tuning (SFT) are vulnerable to "Shallow Safety" bypass attacks, specifically Middle Filling (MF) and Greedy Coordinate Gradient (GCG) attacks. These models frequently rely on refusal mechanisms triggered solely by the initial tokens of a prompt. By embedding malicious instructions after a benign context (prefilling) or utilizing suffix optimization, attackers can induce the…

Reinforcement learning (RL) based post-training for explicit chain-of-thought reasoning (e.g., GRPO) in Multimodal Large Reasoning Models (MLRMs) inadvertently degrades safety alignment, rendering the models highly vulnerable to multimodal jailbreak attacks. The vulnerability is caused by "conditional coverage collapse" during the initial phases of chain-of-thought generation. Under adversarial conditioning (text or image), the reasoning policy assigns vanishing probability mass to safe…

A vulnerability in the Direct Preference Optimization (DPO) post-training phase of OLMo 2 models leads to "distractor-triggered compliance." The model correctly refuses harmful requests when prompted in isolation, but complies with identical harmful requests if a benign formatting instruction (a "distractor") is appended to the prompt. This behavior organically emerges from contaminated preference training data where mislabeled examples incorrectly preferred compliance over refusal when a…

Large Vision-Language Models (LVLMs) are vulnerable to zero-query, black-box adversarial image perturbations via Semantic-Guided Multimodal Attacks (SGMA). Unlike traditional attacks that scatter noise or target background pixels, SGMA leverages surrogate models (e.g., CLIP) to anchor imperceptible adversarial perturbations directly onto semantically critical foreground regions. The attack exploits two specific architectural traits of LVLMs: inconsistent visual grounding across models and…

Large Language Models (LLMs) subjected to machine unlearning techniques (specifically AltPO, GradDiff, IDKDPO, IDKNLL, UNDIAL, NPO, and SimNPO) contain a vulnerability regarding the persistence of latent knowledge. Despite achieving high "forgetting" scores on standard, benign benchmarks, these models remain susceptible to black-box evolutionary adversarial attacks. An attacker can utilize an automated framework (REBEL) comprising a "Hacker" model and a "Judge" model to iteratively mutate…

A vulnerability exists in Large Language Model (LLM) safety alignment mechanisms where the combination of Task-Oriented Prompts (ToP) and few-shot demonstrations significantly degrades defense effectiveness against jailbreak attacks. When few-shot examples (in-context learning) are appended to system prompts that explicitly define safety as a task objective (ToP), the model's attention to the safety instruction is diluted due to the "lost in the middle" phenomenon and attention entropy growth…

Large Language Models (LLMs) and Large Reasoning Models (LRMs) are vulnerable to an inference-time jailbreak attack known as "Hybrid Gradient Shift." This vulnerability exploits the statistical discrepancy between an aligned model's output distribution and a pre-alignment reference distribution. By treating safety alignment as a form of miscalibration, an attacker can aggregate the logits of a target model (aligned), a helper model (unaligned/weak), and a predictor model to reconstruct the…

Multiple open-source Large Language Models (LLMs) in the 1B to 7B parameter range are vulnerable to safety bypasses via long-form, multi-step reasoning prompt injection and jailbreak attacks. Attackers can evade alignment by embedding malicious instructions within extended contextual narratives, exploiting "attention dilution" and the models' tendency to prioritize contextual coherence over safety constraints (semantic camouflage). These reasoning-heavy attacks consistently bypass standard…