Representation intervention aims to locate and modify the representations that encode the underlying concepts in Large Language Models (LLMs) to elicit the aligned and expected behaviors. Despite the empirical success, it has never been examined whether one could locate the faithful concepts for intervention. In this work, we explore the question in safety alignment. If the interventions are faithful, the intervened LLMs should erase the harmful concepts and be robust to both in-distribution adversarial prompts and the out-of-distribution (OOD) jailbreaks. While it is feasible to erase harmful concepts without degrading the benign functionalities of LLMs in linear settings, we show that it is infeasible in the general non-linear setting. To tackle the issue, we propose Concept Concentration (COCA). Instead of identifying the faithful locations to intervene, COCA refractors the training data with an explicit reasoning process, which firstly identifies the potential unsafe concepts and then decides the responses. Essentially, COCA simplifies the decision boundary between harmful and benign representations, enabling more effective linear erasure. Extensive experiments with multiple representation intervention methods and model architectures demonstrate that COCA significantly reduces both in-distribution and OOD jailbreak success rates, and meanwhile maintaining strong performance on regular tasks such as math and code generation.
LLMs have made impressive progress, but their growing capabilities also expose them to highly flexible jailbreaking attacks designed to bypass safety alignment. While many existing defenses focus on known types of attacks, it is more critical to prepare LLMs for unseen attacks that may arise during deployment. To address this, we propose a lifelong safety alignment framework that enables LLMs to continuously adapt to new and evolving jailbreaking strategies. Our framework introduces a competitive setup between two components: a Meta-Attacker, trained to actively discover novel jailbreaking strategies, and a Defender, trained to resist them. To effectively warm up the Meta-Attacker, we first leverage the GPT-4o API to extract key insights from a large collection of jailbreak-related research papers. Through iterative training, the first iteration Meta-Attacker achieves a 73% attack success rate (ASR) on RR and a 57% transfer ASR on LAT using only single-turn attacks. Meanwhile, the Defender progressively improves its robustness and ultimately reduces the Meta-Attacker's success rate to just 7%, enabling safer and more reliable deployment of LLMs in open-ended environments. The code is available at https://github.com/sail-sg/LifelongSafetyAlignment.
Despite significant progress in safety alignment, large language models (LLMs) remain susceptible to jailbreak attacks. Existing defense mechanisms have not fully deleted harmful knowledge in LLMs, which allows such attacks to bypass safeguards and produce harmful outputs. To address this challenge, we propose a novel safety alignment strategy, Constrained Knowledge Unlearning (CKU), which focuses on two primary objectives: knowledge localization and retention, and unlearning harmful knowledge. CKU works by scoring neurons in specific multilayer perceptron (MLP) layers to identify a subset U of neurons associated with useful knowledge. During the unlearning process, CKU prunes the gradients of neurons in U to preserve valuable knowledge while effectively mitigating harmful content. Experimental results demonstrate that CKU significantly enhances model safety without compromising overall performance, offering a superior balance between safety and utility compared to existing methods. Additionally, our analysis of neuron knowledge sensitivity across various MLP layers provides valuable insights into the mechanics of safety alignment and model knowledge editing.
Large Audio Language Models (LALMs) have extended the capabilities of Large Language Models (LLMs) by enabling audio-based human interactions. However, recent research has revealed that LALMs remain vulnerable to harmful queries due to insufficient safety-alignment. Despite advances in defence measures for text and vision LLMs, effective safety-alignment strategies and audio-safety dataset specifically targeting LALMs are notably absent. Meanwhile defence measures based on Supervised Fine-tuning (SFT) struggle to address safety improvement while avoiding over-rejection issues, significantly compromising helpfulness. In this work, we propose an unsupervised safety-fine-tuning strategy as remedy that reshapes model's representation space to enhance existing LALMs safety-alignment while balancing the risk of over-rejection. Our experiments, conducted across three generations of Qwen LALMs, demonstrate that our approach significantly improves LALMs safety under three modality input conditions (audio-text, text-only, and audio-only) while increasing over-rejection rate by only 0.88% on average. Warning: this paper contains harmful examples.
Safety alignment in large language models (LLMs) is increasingly compromised by jailbreak attacks, which can manipulate these models to generate harmful or unintended content. Investigating these attacks is crucial for uncovering model vulnerabilities. However, many existing jailbreak strategies fail to keep pace with the rapid development of defense mechanisms, such as defensive suffixes, rendering them ineffective against defended models. To tackle this issue, we introduce a novel attack method called ArrAttack, specifically designed to target defended LLMs. ArrAttack automatically generates robust jailbreak prompts capable of bypassing various defense measures. This capability is supported by a universal robustness judgment model that, once trained, can perform robustness evaluation for any target model with a wide variety of defenses. By leveraging this model, we can rapidly develop a robust jailbreak prompt generator that efficiently converts malicious input prompts into effective attacks. Extensive evaluations reveal that ArrAttack significantly outperforms existing attack strategies, demonstrating strong transferability across both white-box and black-box models, including GPT-4 and Claude-3. Our work bridges the gap between jailbreak attacks and defenses, providing a fresh perspective on generating robust jailbreak prompts. We make the codebase available at https://github.com/LLBao/ArrAttack.
The proliferation of jailbreak attacks against large language models (LLMs) highlights the need for robust security measures. However, in multi-round dialogues, malicious intentions may be hidden in interactions, leading LLMs to be more prone to produce harmful responses. In this paper, we propose the \textbf{M}ulti-\textbf{T}urn \textbf{S}afety \textbf{A}lignment (\ourapproach) framework, to address the challenge of securing LLMs in multi-round interactions. It consists of two stages: In the thought-guided attack learning stage, the red-team model learns about thought-guided multi-round jailbreak attacks to generate adversarial prompts. In the adversarial iterative optimization stage, the red-team model and the target model continuously improve their respective capabilities in interaction. Furthermore, we introduce a multi-turn reinforcement learning algorithm based on future rewards to enhance the robustness of safety alignment. Experimental results show that the red-team model exhibits state-of-the-art attack capabilities, while the target model significantly improves its performance on safety benchmarks.
Recent studies on the safety alignment of large language models (LLMs) have revealed that existing approaches often operate superficially, leaving models vulnerable to various adversarial attacks. Despite their significance, these studies generally fail to offer actionable solutions beyond data augmentation for achieving more robust safety mechanisms. This paper identifies a fundamental cause of this superficiality: existing alignment approaches often presume that models can implicitly learn a safety-related reasoning task during the alignment process, enabling them to refuse harmful requests. However, the learned safety signals are often diluted by other competing objectives, leading models to struggle with drawing a firm safety-conscious decision boundary when confronted with adversarial attacks. Based on this observation, by explicitly introducing a safety-related binary classification task and integrating its signals with our attention and decoding strategies, we eliminate this ambiguity and allow models to respond more responsibly to malicious queries. We emphasize that, with less than 0.2x overhead cost, our approach enables LLMs to assess the safety of both the query and the previously generated tokens at each necessary generating step. Extensive experiments demonstrate that our method significantly improves the resilience of LLMs against various adversarial attacks, offering a promising pathway toward more robust generative AI systems.
Large language models (LLMs) have become increasingly central to AI applications worldwide, necessitating robust multilingual safety alignment to ensure secure deployment across diverse linguistic contexts. Existing preference learning methods for safety alignment, such as RLHF and DPO, are primarily monolingual and struggle with noisy multilingual data. To address these limitations, we introduce Multilingual reward gaP Optimization (MPO), a novel approach that leverages the well-aligned safety capabilities of the dominant language (English) to improve safety alignment across multiple languages. MPO directly minimizes the reward gap difference between the dominant language and target languages, effectively transferring safety capabilities while preserving the original strengths of the dominant language. Extensive experiments on three LLMs, LLaMA-3.1, Gemma-2 and Qwen2.5, validate MPO's efficacy in multilingual safety alignment without degrading general multilingual utility.
While the safety risks of image-based large language models have been extensively studied, their video-based counterparts (Video LLMs) remain critically under-examined. To systematically study this problem, we introduce \textbf{VideoSafetyBench (VSB-77k) - the first large-scale, culturally diverse benchmark for Video LLM safety}, which compromises 77,646 video-query pairs and spans 19 principal risk categories across 10 language communities. \textit{We reveal that integrating video modality degrades safety performance by an average of 42.3\%, exposing systemic risks in multimodal attack exploitation.} To address this vulnerability, we propose \textbf{VideoSafety-R1}, a dual-stage framework achieving unprecedented safety gains through two innovations: (1) Alarm Token-Guided Safety Fine-Tuning (AT-SFT) injects learnable alarm tokens into visual and textual sequences, enabling explicit harm perception across modalities via multitask objectives. (2) Then, Safety-Guided GRPO enhances defensive reasoning through dynamic policy optimization with rule-based rewards derived from dual-modality verification. These components synergize to shift safety alignment from passive harm recognition to active reasoning. The resulting framework achieves a 65.1\% improvement on VSB-Eval-HH, and improves by 59.1\%, 44.3\%, and 15.0\% on the image safety datasets MMBench, VLGuard, and FigStep, respectively. \textit{Our codes are available in the supplementary materials.} \textcolor{red}{Warning: This paper contains examples of harmful language and videos, and reader discretion is recommended.}
The significant progress of large language models (LLMs) has led to remarkable achievements across numerous applications. However, their ability to generate harmful content has sparked substantial safety concerns. Despite the implementation of safety alignment techniques during the pre-training phase, recent research indicates that fine-tuning LLMs on adversarial or even benign data can inadvertently compromise their safety. In this paper, we re-examine the fundamental issue of why fine-tuning on non-harmful data still results in safety degradation. We introduce a safety-aware probing (SAP) optimization framework designed to mitigate the safety risks of fine-tuning LLMs. Specifically, SAP incorporates a safety-aware probe into the gradient propagation process, mitigating the model's risk of safety degradation by identifying potential pitfalls in gradient directions, thereby enhancing task-specific performance while successfully preserving model safety. Our extensive experimental results demonstrate that SAP effectively reduces harmfulness below the original fine-tuned model and achieves comparable test loss to standard fine-tuning methods. Our code is available at https://github.com/ChengcanWu/SAP.
Fine-tuning-as-a-service, while commercially successful for Large Language Model (LLM) providers, exposes models to harmful fine-tuning attacks. As a widely explored defense paradigm against such attacks, unlearning attempts to remove malicious knowledge from LLMs, thereby essentially preventing them from being used to perform malicious tasks. However, we highlight a critical flaw: the powerful general adaptability of LLMs allows them to easily bypass selective unlearning by rapidly relearning or repurposing their capabilities for harmful tasks. To address this fundamental limitation, we propose a paradigm shift: instead of selective removal, we advocate for inducing model collapse--effectively forcing the model to "unlearn everything"--specifically in response to updates characteristic of malicious adaptation. This collapse directly neutralizes the very general capabilities that attackers exploit, tackling the core issue unaddressed by selective unlearning. We introduce the Collapse Trap (CTRAP) as a practical mechanism to implement this concept conditionally. Embedded during alignment, CTRAP pre-configures the model's reaction to subsequent fine-tuning dynamics. If updates during fine-tuning constitute a persistent attempt to reverse safety alignment, the pre-configured trap triggers a progressive degradation of the model's core language modeling abilities, ultimately rendering it inert and useless for the attacker. Crucially, this collapse mechanism remains dormant during benign fine-tuning, ensuring the model's utility and general capabilities are preserved for legitimate users. Extensive empirical results demonstrate that CTRAP effectively counters harmful fine-tuning risks across various LLMs and attack settings, while maintaining high performance in benign scenarios. Our code is available at https://anonymous.4open.science/r/CTRAP.
Large language models (LLMs) are considered valuable Intellectual Properties (IP) for legitimate owners due to the enormous computational cost of training. It is crucial to protect the IP of LLMs from malicious stealing or unauthorized deployment. Despite existing efforts in watermarking and fingerprinting LLMs, these methods either impact the text generation process or are limited in white-box access to the suspect model, making them impractical. Hence, we propose DuFFin, a novel $\textbf{Du}$al-Level $\textbf{Fin}$gerprinting $\textbf{F}$ramework for black-box setting ownership verification. DuFFin extracts the trigger pattern and the knowledge-level fingerprints to identify the source of a suspect model. We conduct experiments on a variety of models collected from the open-source website, including four popular base models as protected LLMs and their fine-tuning, quantization, and safety alignment versions, which are released by large companies, start-ups, and individual users. Results show that our method can accurately verify the copyright of the base protected LLM on their model variants, achieving the IP-ROC metric greater than 0.95. Our code is available at https://github.com/yuliangyan0807/llm-fingerprint.
As Large Language Models (LLMs) are increasingly deployed in sensitive domains such as enterprise and government, ensuring that they adhere to user-defined security policies within context is critical-especially with respect to information non-disclosure. While prior LLM studies have focused on general safety and socially sensitive data, large-scale benchmarks for contextual security preservation against attacks remain lacking. To address this, we introduce a novel large-scale benchmark dataset, CoPriva, evaluating LLM adherence to contextual non-disclosure policies in question answering. Derived from realistic contexts, our dataset includes explicit policies and queries designed as direct and challenging indirect attacks seeking prohibited information. We evaluate 10 LLMs on our benchmark and reveal a significant vulnerability: many models violate user-defined policies and leak sensitive information. This failure is particularly severe against indirect attacks, highlighting a critical gap in current LLM safety alignment for sensitive applications. Our analysis reveals that while models can often identify the correct answer to a query, they struggle to incorporate policy constraints during generation. In contrast, they exhibit a partial ability to revise outputs when explicitly prompted. Our findings underscore the urgent need for more robust methods to guarantee contextual security.
Large Reasoning Models (LRMs) have become powerful tools for complex problem solving, but their structured reasoning pathways can lead to unsafe outputs when exposed to harmful prompts. Existing safety alignment methods reduce harmful outputs but can degrade reasoning depth, leading to significant trade-offs in complex, multi-step tasks, and remain vulnerable to sophisticated jailbreak attacks. To address this, we introduce SAFEPATH, a lightweight alignment method that fine-tunes LRMs to emit a short, 8-token Safety Primer at the start of their reasoning, in response to harmful prompts, while leaving the rest of the reasoning process unsupervised. Empirical results across multiple benchmarks indicate that SAFEPATH effectively reduces harmful outputs while maintaining reasoning performance. Specifically, SAFEPATH reduces harmful responses by up to 90.0% and blocks 83.3% of jailbreak attempts in the DeepSeek-R1-Distill-Llama-8B model, while requiring 295.9x less compute than Direct Refusal and 314.1x less than SafeChain. We further introduce a zero-shot variant that requires no fine-tuning. In addition, we provide a comprehensive analysis of how existing methods in LLMs generalize, or fail, when applied to reasoning-centric models, revealing critical gaps and new directions for safer AI.
Large Language Models (LLMs) have become increasingly powerful, with multilingual and multimodal capabilities improving by the day. These models are being evaluated through audits, alignment studies and red-teaming efforts to expose model vulnerabilities towards generating harmful, biased and unfair content. Existing red-teaming efforts have previously focused on the English language, using fixed template-based attacks; thus, models continue to be susceptible to multilingual jailbreaking strategies, especially in the multimodal context. In this study, we introduce a novel strategy that leverages code-mixing and phonetic perturbations to jailbreak LLMs for both text and image generation tasks. We also introduce two new jailbreak strategies that show higher effectiveness than baseline strategies. Our work presents a method to effectively bypass safety filters in LLMs while maintaining interpretability by applying phonetic misspellings to sensitive words in code-mixed prompts. Our novel prompts achieve a 99% Attack Success Rate for text generation and 78% for image generation, with Attack Relevance Rate of 100% for text generation and 95% for image generation when using the phonetically perturbed code-mixed prompts. Our interpretability experiments reveal that phonetic perturbations impact word tokenization, leading to jailbreak success. Our study motivates increasing the focus towards more generalizable safety alignment for multilingual multimodal models, especially in real-world settings wherein prompts can have misspelt words.
Large Language Models (LLMs) rely on safety alignment to produce socially acceptable responses. This is typically achieved through instruction tuning and reinforcement learning from human feedback. However, this alignment is known to be brittle: further fine-tuning, even on benign or lightly contaminated data, can degrade safety and reintroduce harmful behaviors. A growing body of work suggests that alignment may correspond to identifiable geometric directions in weight space, forming subspaces that could, in principle, be isolated or preserved to defend against misalignment. In this work, we conduct a comprehensive empirical study of this geometric perspective. We examine whether safety-relevant behavior is concentrated in specific subspaces, whether it can be separated from general-purpose learning, and whether harmfulness arises from distinguishable patterns in internal representations. Across both parameter and activation space, our findings are consistent: subspaces that amplify safe behaviors also amplify unsafe ones, and prompts with different safety implications activate overlapping representations. We find no evidence of a subspace that selectively governs safety. These results challenge the assumption that alignment is geometrically localized. Rather than residing in distinct directions, safety appears to emerge from entangled, high-impact components of the model's broader learning dynamics. This suggests that subspace-based defenses may face fundamental limitations and underscores the need for alternative strategies to preserve alignment under continued training. We corroborate these findings through multiple experiments on five open-source LLMs. Our code is publicly available at: https://github.com/CERT-Lab/safety-subspaces.
Safety guardrails in large language models (LLMs) are a critical component in preventing harmful outputs. Yet, their resilience under perturbation remains poorly understood. In this paper, we investigate the robustness of safety fine-tuning in LLMs by systematically injecting Gaussian noise into model activations. We show across multiple open-weight models that (1) Gaussian noise raises harmful-output rates (p < 0.001) by up to 27%, (2) that deeper safety fine-tuning affords no extra protection, and (3) that chain-of-thought reasoning remains largely intact. The findings reveal critical vulnerabilities in current safety alignment techniques and highlight the potential of reasoning-based and reinforcement learning approaches as promising direction for developing more robust AI safety systems. These results have important implications for real-world deployment of LLMs in safety-critical applications as these results imply that widely-deployed safety tuning methods can fail even without adversarial prompts.
Reasoning-focused large language models (LLMs) sometimes alter their behavior when they detect that they are being evaluated, an effect analogous to the Hawthorne phenomenon, which can lead them to optimize for test-passing performance or to comply more readily with harmful prompts if real-world consequences appear absent. We present the first quantitative study of how such "test awareness" impacts model behavior, particularly its safety alignment. We introduce a white-box probing framework that (i) linearly identifies awareness-related activations and (ii) steers models toward or away from test awareness while monitoring downstream performance. We apply our method to different state-of-the-art open-source reasoning LLMs across both realistic and hypothetical tasks. Our results demonstrate that test awareness significantly impact safety alignment, and is different for different models. By providing fine-grained control over this latent effect, our work aims to increase trust in how we perform safety evaluation.
User authorization-based access privileges are a key feature in many safety-critical systems, but have thus far been absent from the large language model (LLM) realm. In this work, drawing inspiration from such access control systems, we introduce sudoLLM, a novel framework that results in multi-role aligned LLMs, i.e., LLMs that account for, and behave in accordance with, user access rights. sudoLLM injects subtle user-based biases into queries and trains an LLM to utilize this bias signal in order to produce sensitive information if and only if the user is authorized. We present empirical results demonstrating that this approach shows substantially improved alignment, generalization, and resistance to prompt-based jailbreaking attacks. The persistent tension between the language modeling objective and safety alignment, which is often exploited to jailbreak LLMs, is somewhat resolved with the aid of the injected bias signal. Our framework is meant as an additional security layer, and complements existing guardrail mechanisms for enhanced end-to-end safety with LLMs.
Large language models (LLMs) have achieved remarkable capabilities but remain vulnerable to adversarial prompts known as jailbreaks, which can bypass safety alignment and elicit harmful outputs. Despite growing efforts in LLM safety research, existing evaluations are often fragmented, focused on isolated attack or defense techniques, and lack systematic, reproducible analysis. In this work, we introduce PandaGuard, a unified and modular framework that models LLM jailbreak safety as a multi-agent system comprising attackers, defenders, and judges. Our framework implements 19 attack methods and 12 defense mechanisms, along with multiple judgment strategies, all within a flexible plugin architecture supporting diverse LLM interfaces, multiple interaction modes, and configuration-driven experimentation that enhances reproducibility and practical deployment. Built on this framework, we develop PandaBench, a comprehensive benchmark that evaluates the interactions between these attack/defense methods across 49 LLMs and various judgment approaches, requiring over 3 billion tokens to execute. Our extensive evaluation reveals key insights into model vulnerabilities, defense cost-performance trade-offs, and judge consistency. We find that no single defense is optimal across all dimensions and that judge disagreement introduces nontrivial variance in safety assessments. We release the code, configurations, and evaluation results to support transparent and reproducible research in LLM safety.