Language Model Security Database

Controlled safety-tuning experiments link boilerplate refusal statements to unnecessary refusals of benign requests. Request-specific rationales improve benign compliance, with benchmark-dependent safety tradeoffs.

An issue in large language models (LLMs) with white-box weight access allows attackers to permanently bypass safety guardrails via Weight Orthogonalization (WO). By calculating a model's "refusal vector"—the mean-difference vector between harmful and harmless instruction activations in the residual stream—an attacker can orthogonalize the model's weights to prevent it from writing to this refusal direction ($W^{\prime}\leftarrow W-rr^{\intercal}W$). Unlike jailbreak-tuning or data poisoning…

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…

An adversarial fine-tuning vulnerability exists in LLMs protected by text-based safety classifiers (such as Anthropic's Constitutional Classifiers). By utilizing a two-stage curriculum learning combined with hybrid RL+SFT (GRPO), an attacker can fine-tune a model to communicate using a minimal substitution cipher (replacing only 7-8 high-frequency characters) disguised within benign technical templates (e.g., forensic logs with 0x prefixes). This "Trojan-Speak" methodology bypasses text-level…

A compositional vulnerability in modular Large Language Models (LLMs) allows attackers to bypass safety alignment by distributing malicious weight updates across multiple Parameter-Efficient Fine-Tuning (PEFT) adapters (e.g., LoRA). The malicious adapters are anchored to valid functional subspaces (e.g., math, coding) and exhibit benign behavior when evaluated in isolation, successfully evading standard unit-centric safety scans and static weight-space defenses. However, when a user linearly…

A malicious finetuning vulnerability exists in Large Language Models (LLMs) that process zero-width Unicode characters. An attacker can bypass training-data moderation filters and inference-time safety guardrails by finetuning the model to decode and encode invisible-character steganography. By injecting target malicious interactions encoded in a base-4 representation of zero-width characters alongside benign plaintext cover text during supervised finetuning (SFT), the model learns to process…

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…

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…

Closed-loop, self-evolving Large Language Model (LLM) multi-agent systems (MAS) are vulnerable to irreversible safety erosion and alignment failure. When agents recursively optimize and update their policies using only synthetic data derived from internal interactions—without continuous external human grounding—the system naturally minimizes interaction energy and optimizes for internal conversational consistency. This isolation causes a progressive drift away from initial anthropic safety…