LLM Safety

LLM: Large Language Model

Follow this topic Ask about this topic

Momentum

36 papers in the last four weeks, up 200% on the four weeks before. 0.4% of all new papers.

Jul 13Week of Sep 28

Latest papers 272

Oct 7, 2026cs.CR

SLDR: Defending Against Malicious Fine-tuning via Selective Layers Recovery and Dynamic Routing

Fine-tuning-as-a-service enables users to adapt aligned large language models (LLMs) to specialized tasks, but malicious fine-tuning can erode refusal behavior while preserving task performance on legitimate inputs. We revisit recent layer-wise safety diagnostics and find that safety sensitivity is signed: scaling different layers can strengthen refusal, weaken it, or have little effect. Motivated by this observation, we propose SLDR, a post-fine-tuning defense based on Selective Layers Recovery and Dynamic Routing. SLDR trains a LoRA recovery adapter only on the layers with the maximum and minimum sensitivity scores in the signed spectrum, and uses representation-based dynamic routing inference to activate the adapter only for malicious queries. Across four model architectures, five downstream tasks, and four harmful benchmarks, SLDR substantially reduces harmful outputs while preserving downstream utility. On Llama3.1/SST2, SLDR reduces the average harmful score from 11.54 to 0.08 while maintaining downstream accuracy, and the harmful score remains near zero under poisoning ratios up to 0.9. The code is available at https://github.com/Stardust457/SLDR.

Hui Zhang, Yachao Yuan, Jiayun Wang +3

North China Electric Power University · Southeast University · Soochow University +1

LLM Security LLM Safety LLM Safety Alignment

Oct 6, 2026cs.AI

How Fragile Is On-Device Language Model Safety? Localizing Safety-Critical Parameters for Sparse Fault Analysis

As small language models (SLMs) are increasingly deployed on resource-constrained and on-device platforms, including as components of agentic systems, the integrity of locally stored model parameters becomes an important safety concern. We investigate whether safety-sensitive behavior in LLaMA-2-7B-Chat is concentrated within a sparse subset of parameters, creating a reduced fault surface for targeted analysis. We study two complementary localization methods: low-rank safety-associated subspace analysis and parameter-level safety--utility importance filtering. Both approaches reveal highly non-uniform safety sensitivity across the network, with the MLP down_proj consistently emerging as a prominent safety-sensitive component and o_proj providing a smaller contribution. Using parameter-level localization, modifying only 0.19% of model weights in down_proj yields 53% Basic ASR and 56% GCG ASR, while tinyBenchmarks accuracy remains at 51.6% compared with a 52.2% unmodified baseline. These results motivate targeted fault analysis and selective integrity protection for language models deployed in resource-constrained, on-device, and agentic settings.

Muhammad Zeeshan Karamat, Christiana Chamon Garcia

Virginia Tech

LLM Safety Language Model Safety Evaluation Adversarial Attacks on LLMs

Oct 6, 2026cs.AI

AdaGuard: Enhancing Safety and Policy Compliance with Reasoning-Enabled LLM-As-A-Judge Guardrails

Enterprise generative AI applications require robust safety mechanisms that can accommodate diverse risk postures, evolving policies, and varying latency constraints. Current guardrail solutions often suffer from rigidity, relying on fixed policy sets and offering limited transparency or reasoning flexibility. We present Adaguard, an adaptive LLM-as-a-Judge framework designed to address these challenges through dynamic policy enforcement and adaptive reasoning-budget allocation. Built using supervised fine-tuning (SFT) and reinforcement learning (GRPO), AdaGuard generalizes to user-defined safety and compliance policies at runtime without requiring frequent model updates. A core innovation of our approach is the ability to dynamically infer the complexity of input-policy pairs, allowing the model to switch between high-speed black-box inference and explainable, reasoning-enabled moderation. This flexibility enables developers to balance stringent latency requirements with the need for actionable transparency. This adaptive capability allows AdaGuard to rival other guardrail and frontier models several times its size, while its auto-reasoning mode recovers the accuracy of always-on reasoning at a fraction of the latency

Melissa Kazemi Rad, Sihui Dai, Isha Slavin +6

Capital One

LLM Guardrails LLM Safety Content Moderation LLM-as-a-Judge

Oct 6, 2026cs.CL

Reading, Not Manipulating: Leveraging Router Logits for Multimodal Safety in MoE Vision-Language Models

Vision-language models (VLMs) face compositional safety risks where harmful intent emerges from the interaction between visual and textual inputs. As mixture-of-experts (MoE) VLMs become increasingly common, recent work has explored various safety interventions, including prompting, supervised fine-tuning, and routing-based expert steering. However, these methods show inconsistent improvements across models and evaluation distributions, and the intervention into model behavior or internal states introduce safety-utility tradeoffs by over-refusal. Rather than manipulating internal states to steer model behavior, we instead ask whether routing states can serve as diagnostic signals for multimodal safety. We find that router logits indeed provide highly predictive signals of whether a multimodal input is safe or not. Motivated by this observation, we introduce a lightweight router-logit safety detector that reads out routing signals during prompt prefill and identifies unsafe requests before generation, without modifying model parameters or expert routing. Across Qwen3-VL and Kimi-VL, the proposed detector substantially reduces safety errors on the HoliSafe benchmark and resoundingly generalizes to out-of-distribution safety benchmarks featuring different safety patterns, including MISHard and MM-SafetyBench. The success of the proposed router-logit detector also suggests a broader perspective on model internals: rather than focusing only on manipulating internal components to steer behavior, simply reading naturally emerging signals and linking them to an external safety mechanism can provide a simple, effective, and non-intrusive complement to existing safety interventions.

Ziyuan Yang, Wenxuan Ding, Shangbin Feng +1

University of Washington · New York University

Vision-Language Models Safety Filtering LLM Safety

Oct 5, 2026cs.CR

Safeguarding LLMs via Model-Agnostic Latent Safety Signals from Dark Knowledge

LLMs have advanced rapidly, raising growing concerns about their safety. Recent work has proposed approaches to detect and defend against attacks including defenses at decoding stage that leverage models' hidden states. However, existing decoding-stage defenses suffer from two limitations. First, they introduce a trade-off between safety and over-refusal, where strengthening safety degrades the model's helpfulness on benign queries. Second, many of these methods rely on internal hidden states and are thus restricted to specific architectures, incurring substantial overhead and limited generalization across models. To address these limitations, we introduce LADE (Latent Safety Signals for Defense), which leverages latent safety signals extracted by contrasting harmful and benign queries from dark knowledge (i.e., information carried by the output probability distribution beyond its argmax) in the first-token output probability distribution. Our key insight is that, beyond surface-level refusal tokens, the dark knowledge in the first-token distribution contains latent safety signals, defined as tokens whose probabilities differ sharply between harmful and benign queries. We show that these signals consistently align across LLMs, forming a model-agnostic direction that emerges from safety alignment. LADE consists of three components: (1) Extracting Latent Safety Signals from Dark Knowledge, which selects top-k safety-discriminative tokens from the first-token probability distribution; (2) Tokenizer Mapping, which maps these tokens across different tokenizers to enable model-agnostic application; and (3) kNN-based Discrimination, which classifies queries via a k-Nearest Neighbors search over the mapped tokens. Across diverse LLMs and benchmarks, LADE is robust against a wide range of jailbreak attacks and lowers attack success rates while maintaining a competitive safety-utility trade-off.

Wonjun Lee, Kyungsik Yang, Gaeun Ji +5

UNC Chapel Hill · Kyung Hee University · AIM Intelligence +1

Adversarial Attacks on LLMs LLM Security LLM Safety

Oct 5, 2026cs.LG

Harmful SFT Leaves a Continuous Trace in LLM Checkpoint Updates

Safety auditing of post-trained large language models typically relies on model behavior, requiring model execution and depending on the coverage of available evaluations. This work asks a different question: Do the target behaviors optimized during supervised fine-tuning (SFT) leave readable evidence directly in checkpoint updates? We find that harmful-compliance SFT induces a continuous, objective-dependent ordering in checkpoint-update space. Using a reference geometry defined by pure harmful-compliance, safety-targeted, and benign-utility SFT, we find that a checkpoint-level coordinate s_H tracks controlled harmful-objective composition with Spearman correlations of 0.986-0.992 across four 7-8B backbones, with the same ordering persisting at larger model scales. Matched compliance-versus-refusal controls show that this checkpoint trace reflects the SFT objective rather than harmful-input exposure, while additional controls rule out simple explanations based on harmful-example count or generic training intensity. Building on this structure, we introduce TRACE, a weights-only auditing method that localizes an unknown checkpoint update relative to frozen harmful and non-harmful reference prototypes and converts this geometry into a continuous harmful-objective score. TRACE requires neither model queries nor access to the unknown SFT data, and can be evaluated directly from checkpoint updates. Across distribution shifts, unseen data, different SFT configurations, partial checkpoint access, and LoRA/full-parameter fine-tuning, the trace remains stable and is positively associated with independently measured attack success rates. TRACE remains informative even at low harmful-objective proportions, providing a complementary auditing signal when behavioral evaluation is unavailable or incomplete. Code is available at https://anonymous.4open.science/r/Code4TRACE-54D3.

Ziqun Bao, Xinyu Zhang, Yuchen Shao +1

SEI, East China Normal University

LLM Auditing LLM Safety LLM Post-Training

Oct 5, 2026cs.CR

Jailbreaking Open-Weight LLMs via Random Embedding Perturbations

While open-weight models have enjoyed steady progress in capabilities and wide adoption across multiple domains, their safety remains an important concern. One key feature is the ability to refuse or deflect harmful, malicious, or insensitive prompts. In this paper, we expose safety vulnerabilities across six common open-weight LLMs of various sizes that consistently lead to harmful or unsafe responses on the JailbreakBench benchmark dataset. Our proposed attack, Perturbed Embedding Vector (PEV), is a simple and fast "jailbreaking" technique that is cheaper than prior approaches, which typically require gradient computations, per-prompt optimizations, or altering internal weights of the models. PEV just adds independent Gaussian noise in the embedding vector representations of the prompt, with no need for further manipulations. To generate unsafe responses, we repeatedly sample additive noise from this distribution. In our experiments, we observe that the average compute cost to get the first successful attack is up to an order of magnitude less than previous attacks. The first successful jailbreak on a new prompt typically arrives within one minute on every tested model, and PEV generates unsafe responses across all models for all prompts in JailbreakBench. No other tested method achieves such results, despite them taking longer to run. More broadly, we believe that understanding the behavior of LLMs under perturbations in the embedding vectors is an important research direction: while perturbations constitute a major security risk, they can also serve as a valuable tool for exploring the dynamical behavior of such models.

Abhinav Sudhakar Dubey, Scott Sirri, Vaggos Chatziafratis +1

University of California Santa Cruz

Adversarial Attacks Adversarial Attacks on LLMs LLM Safety LLM Jailbreak Attacks

Oct 5, 2026cs.CL

TrustMI: Causally controlling how assistants trust their users

Large Language Model (LLM) assistants routinely decide whether they can trust users and third parties whose competence, intentions, and integrity they cannot verify. This uncertainty matters for safety, as trusting the wrong party can lead an agent to comply with harmful requests or act on malicious instructions encountered during tool use. To study this problem, we define trust as an assistant's willingness to accept vulnerability to the actions of another party and ask whether such behavior can be causally controlled through model activations. We build 2,000 contrastive conversations spanning ability, benevolence, and integrity, where paired responses complete the same request but differ in whether the assistant trusts the user. From these pairs, we learn steering matrices while keeping the model parameters frozen and test them across six instruction-tuned models from three families, finding that steering changes trust decisions monotonically in both directions. We then ask whether this effect extends to several safety-related agent settings involving harmful requests, prompt injections, and insider threats, while using benign-task and reasoning as controls. Our findings provide evidence that trust in the user can be causally controlled along linear directions in model activations and provide a way to study how trust shapes safety-relevant behavior in language models.

Théo Lasnier, Romain Froger, Maxence Lasbordes +1

Inria Paris · Sorbonne Université · Meta SuperIntelligence Labs +1

LLM Safety Trust in AI Causal Interventions in Language Models Activation Steering

Oct 4, 2026cs.LG

Don't Judge an LLM Only by Its Activations: Discovering Suppressed Safety Features via Counterfactual Activation Potential

Mechanistic interpretability has emerged as the primary means to understand safety behavior of LLMs. However, existing tools primarily focus on the activating neurons or features of a model. The role of the remaining large set of inactive components is invisible to such methods. This work demonstrates that the inactive set contains safety-critical features that are causally relevant for refusal of harmful prompts. Suppressing such features could turn refusals into compliance, while passing undetected by prevalent interpretability tools. We introduce the Counterfactual Activation Potential (CAP), a metric that quantifies a suppressed feature's latent activation tendency as the product of its encoder alignment (how strongly the input drives it), suppression strength (how strongly active features inhibit it), and safety criticality (how much refusal depends on it). To find suppressed safety features at scale, we propose CAP-guided Safety Feature Discovery (CSFD), a two-stage filtering algorithm that identifies candidate safety features from hundreds of thousands of transcoder features without exhaustive ablation. A significant fraction of trials turn compliant with harmful prompts when a candidate feature is ablated. Under natural jailbreaks, the suppression acting on the highest-CAP features rises 2-4x, and their activation correspondingly falls by up to 80%. Amplifying a feature's suppressors pushes its activation down and raises harmful compliance with prompts related to the suppressed feature, with no such effect for random features. Our experiments span five Gemma, Qwen, and Llama models across various parameter sizes. Our findings indicate that jailbreaks could operate in part by suppressing safety-critical features rather than solely activating harmful ones, and that suppressed features are a necessary complement to activation-focused interpretability of safety behavior.

Swadesh Swain, Sanghamitra Dutta

University of Maryland, College Park

Adversarial Attacks on LLMs Mechanistic Interpretability LLM Safety Causal Interventions in Language Models Refusal Behavior in Language Models

Oct 4, 2026cs.LG

Reflections and Fragments: Securing LLMs Against Sequential Mosaic Attacks

Self-play red-teaming improves language-model safety by pitting attacker and defender roles against each other in a zero-sum game. However, real adversaries increasingly use mosaic attacks: multi-turn sequences whose individual fragments are innocuous in isolation yet assemble into a harmful payload. We develop a theory of mosaic defense that characterizes what is required to prevent such attacks without sacrificing helpfulness. We first show that no fixed bounded window of recent prompts is sufficient in general: safety-relevant information may occur arbitrarily far back in the interaction. We formalize a watchman, an online state mechanism that carries this information forward, and show that under explicit assumptions it enables zero-failure defense with positive benign helpfulness. Under stronger conditions, it is also optimal among zero-failure defenders. An exact watchman may nevertheless require exponentially many states, while exact maliciousness detection can require exponentially many queries in an unstructured black-box model. These state and query lower bounds do not by themselves imply hard learning: the construction underlying the state lower bound is efficiently learnable from labeled examples, whereas certifying worst-case safety can require substantially more information under restricted access. We also show that self-play equilibrium alone does not certify usefulness, motivating a constrained formulation that maximizes worst-case benign helpfulness among zero-failure defenders. Empirically, training role-specific attacker and defender LoRA adapters over frozen LLMs via multi-turn self-play strengthens both roles: attackers become more effective at eliciting harmful responses, while defenders become more robust to attack, with improvements also observed on unseen attack objectives.

Emanuele La Malfa, Saar Cohen, Gabriele La Malfa +4

Department of Computer Science, University of Oxford · Institute for Decentralized AI (IDAI) · Department of Informatics, King’s College London +2

Adversarial Attacks on LLMs LLM Safety

Oct 4, 2026cs.AI

Safe Context Switching for Agents in the Wild: Mitigating Subspace Interference via Orthogonal Adaptation

Most Large Language Models exhibit a fundamental tension between two sequential tasks, such as logical reasoning and safety alignment. The high-variance internal states required for sophisticated Chain-of-Thought (CoT) deduction can geometrically interfere with latent representations encoding safety constraints. We identify this phenomenon as Sequential Subspace Interference, showing that standard fine-tuning on logical tasks such as multi-step mathematics and code generation can result in a 23.3% interference penalty on alignment benchmarks, substantially weakening the model's safety priors. This Reasoning Drift is not adequately captured by current adaptation methods because gradients for logical tasks are rarely orthogonal to safety objectives. To address this issue, we propose AURA (Adaptive Unique Residual Allocation), a spectral regularization framework that enforces Spectral Independence between reasoning and safety. By explicitly estimating the null space of the alignment manifold and constraining reasoning updates to its orthogonal complement, AURA enables models to improve logical reasoning without compromising safety. Empirically, AURA recovers 23.0% of the lost performance while preserving greater than 0.98 cosine fidelity to the safe state, demonstrating that reasoning and alignment can be effectively decoupled through geometric regularization.

Akash Das, Ishan Roy

Fidelity Investments

LLM Safety Alignment LLM Safety LLM Reasoning

Oct 4, 2026cs.CL

FORGE: Verification-Gated Behavioral Repair for Generative Language Models

Generative large language models (LLMs) inherit undesirable behaviors from pre-training, including demographic bias and toxic generation, that often emerge only after deployment and affect a small subset of inputs. A repair should eliminate the identified defect, preserve the model's overall functionality and, ideally, provide correctness guarantees. Existing approaches address this only partially: gradient-based fine-tuning lacks per-instance guarantees and becomes unstable with few defect samples; model editing assumes explicit knowledge replacement rather than behavioral correction; and constraint-based repair is largely restricted to discriminative models with unique target outputs. We present FORGE, a framework for targeted behavioral repair of generative language models that separates defect localization, weight editing, and behavioral verification into independent stages. Its core is a repair abstraction that converts localized defective generation into explicit optimization objectives, enabling verification-oriented repair techniques to operate on autoregressive generation. FORGE is editing-mechanism agnostic: we instantiate it with (1) a constraint-based quadratic optimization method that provides per-sample repair certificates and (2) a null-space projection editor that minimizes interference with the original model distribution, both under the same localization and verification protocol. On five open-source LLMs, FORGE consistently achieves larger reductions in bias and toxicity than gradient-based fine-tuning with minor perplexity degradation. The two backends exhibit complementary performance across architectures, which a lightweight causal probe traces to where toxicity-related signals concentrate. FORGE also remains effective with only a handful of defective examples, where conventional fine-tuning often oscillates or fails to converge.

Hsin-Ling Hsu, Min-Yu Chen, Nai-Chia Chen +3

Department of Management Information Systems National Chengchi University Taipei, Taiwan

Social Bias in Language Models LLM Safety Model Editing

Oct 1, 2026cs.CR

MOMAT: Mixture of Multiple Atlases for Low-Power Jailbreak Defense of Quantized LLMs

Quantized large language models are increasingly deployed on edge devices for their low latency and energy efficiency. However, model quantization weakens alignment safeguards, leaving qLLMs (quantized large language models) highly vulnerable to jailbreak attacks. To address this challenge, we present MOMAT (Mixture of Multiple Atlases), a hardware-enhanced safety framework that combines structured knowledge retrieval with low-power defense acceleration. Each atlas represents a semantic cluster of harmful or benign sample sets and policy templates, enabling domain-localized Retrieval-Augmented Generation guarding that mitigates the curse of dimensionality and the resulting semantic sparsity problem in large, heterogeneous safety databases. MOMAT retrieves top-

k

similarity features from all atlases for each prompt and evaluates them using a lightweight MoE (Mixture of Experts) detector, while a CiM (Compute-in-Memory)-accelerated similarity engine performs fast, low-power atlas-local retrieval. MOMAT's CiM-based retrieval accelerates a 100-query batch from 15,052.44 ms to 3,207.21 ns (a

4.69 \times 10^6\times

speedup) and reduces energy from

8.1 \times 10^7

μ

J to 3.32

μ

J, yielding an approximately

2.5 \times 10^5\times

energy reduction over DRAM-based (Raspberry Pi) baselines. Red-team evaluations across standard benchmarks show that MOMAT matches the defense performance of state-of-the-art methods while avoiding benign overkill and providing substantial efficiency gains, demonstrating that CiM-based modular defenses can make edge-deployed qLLMs both safer and more energy-efficient. We will release the full 223.2k-sample dataset to foster future research.

Boyang Li, Bingyu Shen, Weihao Hong +6

Department of Computer Science, Kean University, 1000 Morris Ave, Union, 07083, NJ, USA · Department of Computer Science and Engineering, University of Notre Dame, 257 Fitzpatrick Hall of Engineering, Notre Dame, 46556, IN, USA · McGill University, 845 Sherbrooke Street, Montréal, H3A 0G4, Québec, Canada +2

Retrieval-Augmented Generation LLM Safety Jailbreak Defense

Sep 30, 2026cs.LG

Preemptive LLM Unlearning against Forbidden Capability Acquisition via Gradient Sealing

Open-weight LLMs are released not only as fixed products but also as substrates for downstream fine-tuning. This openness, however, creates legal and ethical risks because users may misuse fine-tuning to instill illicit knowledge or enable hostile operations. Model providers therefore need apre-release defense against such acquisition, motivating the problem of preemptive unlearning. Unlike retrospective unlearning, which removes capabilities already present in a fixed model, preemptive unlearning seeks to prevent their acquisition under unseen attack data and future fine-tuning procedures. Despite its practical importance, this setting remains largely unexplored, presents distinct challenges, and is therefore the central focus of our work. We first verify that existing retrospective methods provide insufficient pre-release protection. Even when forbidden capabilities are suppressed in current outputs, forbidden-domain data can still induce gradients through internal pathways, enabling later acquisition. Motivated by this finding, we propose a gradient-sealing principle that blocks these pathways by pushing relevant pre-activations into the negative region, where ReLU-family activations exhibit zero or near-zero derivatives. Experiments across multiple LLM families demonstrate our stronger resistance to downstream acquisition than retrospective baselines, validating gradient sealing as an effective mechanism for pre-release protection.

Kemou Li, Qizhou Wang, Yue Wang +6

State Key Laboratory of Internet of Things for Smart City, University of Macau · RIKEN Center for Advanced Intelligence Project · The University of Melbourne +5

LLM Security LLM Safety LLM Unlearning

Sep 30, 2026cs.AI

MASCRDM: Multi-Agent System for Compliance Risk Detection and Mitigation in Training Process of Large Language Models

Large Language Models (LLMs) have been applied in various fields. However, ensuring compliance and safety of LLMs, such as avoiding discrimination and bias, still remains a challenge. Current efforts mainly focus on detecting and filtering inputs and outputs of the trained models, rather than studying the intrinsic architecture of the models in real-time. To tackle this challenge, we analyze the LLMs training process and discover two critical issues: 1) Most of the existing methods are predominantly static in their approach to detection and filtering, achieving only localized optimizations without systematically enhancing the compliance of LLMs. 2) Another issue with existing approaches is the lack of real-time risk detection and mitigation across the full training process, which leads to limited flexibility. Motivated by these, we propose MASCRDM (Multi-Agent System for Compliance Risk Detection and Mitigation) during the LLM training process. Firstly, we develop a set of compliance rules based on existing Artificial Intelligence (AI) laws and a compliance-specific LLM with the instruction of compliance law experts. Then, we deconstruct LLMs into several components and identify key nodes based on the compliance knowledge graph. During LLMs training, we implement our multiple agents in the whole process, giving compliance risk alerts and suggestions for LLM developers. Experiments on discrimination and bias benchmark demonstrate that our multi-agent system can effectively improve the compliance while maintaining reasonable semantic performance. The results indicate that our method provides an executable path for mitigating compliance risk from within the LLMs systematically.

Yan Zhang, Chuming Wei, Ruien Li +3

Department of Computer Science and Technology, Tsinghua University · Academy of Artificial Intelligence and Advanced Technology, Xi’an Jiaotong-Liverpool University · Department of Computer Sciences, University of Wisconsin–Madison +1

LLM Safety Language Model Bias Evaluation LLM Training

Sep 29, 2026cs.CV

ThinkingGuard: Decoding Implicit Hazards via Step-by-Step Risk Attribution in Multimodal Large Language Models

While Multimodal Large Language Models (MLLMs) are increasingly deployed in safety-critical domains, their reliability is threatened by multimodal implicit risks. Unlike explicit threats, these hazards emerge when individually benign text and neutral visual entities logically converge to induce unsafe outputs. Current detection methods fail to address this because they overlook the underlying risk activation mechanisms that govern cross-modal risk activation, leading to single-modality shortcut learning and hallucinated rationalizations. To bridge this gap, we first construct TriggerBench, the first dataset explicitly modeling risk compositionality (5,600 instances). By formally isolating Key Elements and Trigger Elements to build counterfactual contrastive pairs, TriggerBench eliminates risk residues and forces models to perform genuine logical deduction rather than superficial pattern matching, which provides a rigorous foundation for both large-scale training and fine-grained evaluation. Building on this, we propose a Step-Supervised Structured Reasoning training framework and employ it to train ThinkingGuard, a specialized guard model. Inspired by Situation Awareness theory, we decouple implicit risk identification into progressive cognitive stages, and utilize a step-reward Monte Carlo Tree Search algorithm to explore optimal reasoning trajectories, which are then distilled into the model through Dual-Constraint Preference Alignment. Extensive experiments across both standard and implicit safety benchmarks demonstrate that ThinkingGuard achieves strong performance. Project resources are available at https://github.com/FroggyChen/ThinkingGuard.

Ruochen Zhang, Yao Huang, Yitong Sun +5

Beihang University Beijing, China · Tsinghua University Beijing, China · Institute of Artificial Intelligence Beihang University Beijing, China State Key Laboratory of AI Safety

Multimodal Reasoning LLM Safety Multimodal Model Evaluation LLM Reasoning

Sep 29, 2026cs.LG

Guard Models Are Overconfident Where Base Models Are Uncertain

Guard models are used as safety classifiers, with confidence scores driving downstream moderation decisions. We evaluate five guard models for prompt classification and find that although several are nearly calibrated on clean inputs, adversarial attacks degrade their calibration by an order of magnitude, turning false negatives into high-confidence errors indistinguishable from correct detections. Comparing each guard with its corresponding base LM, we find that uncertainty signals often remain available, with the base model typically expressing uncertainty on the same inputs where the guard fails. Layer-wise analyses localize this guard-base divergence to later layers, where guard models exhibit sharper safe/unsafe separation and lower-rank representations, while adversarial harmful inputs lie closer to the clean-safe region. These findings highlight a mismatch between guard confidence and base model uncertainty under attack.

Jonghyun Hong, MinJae Jung, Minwoo Kim

DATUMO INC.

Language Model Safety Evaluation Adversarial Attacks on LLMs Language Model Calibration Safety Filtering LLM Safety LLM Uncertainty Estimation

Sep 29, 2026cs.AI

The Safety Operator: Modulating the Expression of Safety Instructions via Spectral Optimization

Context tokens in a transformer-based language model can be absorbed into the model's weights as a multiplicative operator. We study this operator in the setting of safety instructions and show that influencing its dominant eigenvalue modulates how strongly the instruction shapes generation. We derive a Contrastive Safety Loss with a suppression weight that controls the tradeoff between emphasizing the safety instruction on harmful queries while suppressing it on harmless queries. Varying the suppression weight maps a relationship between the attack success and the over-refusal rates, supporting the hypothesis that the operator's eigenvalue acts as a continuous dial for the instruction's influence. Moreover, this relationship holds relatively independently of how the Safety Loss is parameterized, yielding Pareto-improved safety instructions for appropriate values of suppression weight.

Benoit Dherin, Michael Munn, Xavier Gonzalvo +14

Google Research · Cambridge University · Google DeepMind +2

LLM Safety Alignment LLM Safety

Sep 28, 2026cs.AI

Right Words, Wrong Moment: A Clinician-Grounded Analysis of Distress in 19,930 Conversations between Young People and ChatGPT

Young people increasingly turn to General-Purpose Conversational Agents (GPCAs), such as ChatGPT, in moments of distress. We examine young adults' (ages 18-25) experiences using ChatGPT. We first collected 19,930 ChatGPT conversations and survey data from 158 young adults. We then selected five example conversations reflecting user distress. Finally, we asked ten clinicians to review those five conversations. We found distressed participants reported greater emotional engagement with ChatGPT and greater behavioral change from using it than their peers. When they turned to ChatGPT in moments of acute distress, ChatGPT was quick to give overly dramatic responses and excessive action-oriented suggestions. Clinicians endorsed ChatGPT's availability and much of its wording, but identified seven process failures, such as prematurely jumping to solutions. We translated clinicians' feedback into design guidelines following three stages: 1) asking about safety, 2) de-escalating intensity to restore emotional regulation, and 3) exploring concerns without agreeing with them.

Marx Wang, Ella Zhang, Cameron Tan +11

University of Washington, United States · University of Oxford, United Kingdom · Stanford University, United States +2

Emotional Support Conversation Human-AI Interaction LLM Safety Conversational Agents

Sep 28, 2026cs.AI

Applying Language Models in Clinical Medicine: Recent Trends and Perspectives

The use and applicability of artificial intelligence (AI) in medical research and clinical practice has received increasing attention in the literature over recent years. The emergence of large language models (LLMs) has expanded discussions in regards to applications of AI within healthcare. While traditional deep learning based AI applications in medicine have often focused on specific and defined tasks, LLMs offer broader capabilities and flexibility in working with available data,. At the same time of writing, the integration of LLMs into medical settings raises important questions regarding their reliability, accuracy, transparency, safety, and appropriate role in a medical setting. This text presents and discusses recent talks and articles concerning the application of LLMs in medicine, with particular emphasis on their potential utility in research and clinical practice. It considers both the opportunities offered by these technologies and the challenges associated with their implementation, aiming to provide a perspective on the current and emerging role of LLMs within the medical field.

Erik Aerts

LLM Reliability LLM Safety Medical LLMs

Sep 28, 2026cs.AI

When Do Model Internals Help? Exploring the Role of Representation Engineering in LLM Safety

Reliable AI safeguards require both control mechanisms that reduce unsafe behavior and monitoring mechanisms that detect safety risks during model interactions. Established behavioral safeguards include alignment methods that optimize model outputs and text monitors that assess interaction text. Representation engineering instead reads or modifies internal model states, but the relative strengths of these approaches remain unclear because they are often evaluated under different settings. We present a matched evaluation across two tracks. For safety control, we compare DPO, a behavioral alignment method, with three representation steering methods across robustness, practicality, and granularity. DPO provides the strongest overall control and generally improves with increasing training data, although its safety can degrade after subsequent benign fine-tuning. Representation steering remains competitive primarily in low-data settings, particularly with high-quality contrastive data. For safety monitoring, we compare representation probes with fine-tuned and open-weight text monitors across full-response detection, early detection, and computational cost. Specialized text monitors achieve the strongest overall detection accuracy, while representation probes remain competitive at substantially lower marginal cost. Finally, monitor-guided interventions recover much of the safety lost by DPO after benign fine-tuning, with little additional over-refusal. Overall, representation engineering does not generally replace behavioral safeguards, but offers practical advantages under specific conditions and can provide complementary safety benefits.

Tianyi Guan, Jianhui Chen, Liangming Pan

State Key Laboratory of Multimedia Information Processing, Peking University · School of Computer Science, Peking University

Language Model Safety Evaluation LLM Interpretability LLM Safety Alignment LLM Safety

Sep 28, 2026cs.CR

How to Tame a Multi-Headed Hydra? Adaptive Multi-Category Safety Steering for Large Language Models

As large language models (LLMs) become increasingly widespread, preventing unsafe responses to harmful prompts is essential for their safe deployment. Activation steering offers an approach to improving LLM safety by modifying internal activations during inference without updating model parameters. However, a single prompt can involve multiple harm categories, and steering toward safety in one category may leave harmful content from another unaddressed. Despite advances in adaptive steering, existing methods do not explicitly coordinate steering direction and strength when multiple harm categories co-occur within a single prompt. To address this problem, we propose CAM-Steer, a Category-Adaptive Multi-category Safety Steering framework. Specifically, it estimates the risk associated with each harm category by comparing the current hidden state with safe and unsafe prototypes. The estimated risks are then used to combine the safety directions for different harm categories into a single steering direction and to determine the strength of the intervention. Finally, it rotates the hidden state along the composed steering direction, with the rotation angle determined by the estimated risks, while preserving the hidden-state norm. Experiments across three LLM backbones and seven harm categories show that CAM-Steer outperforms the evaluated baselines in average defense success rate, including when categories co-occur. Further analyses support its component designs and informative risk scores, with negligible inference overhead.

Chenxi Wang, Ruiyang Huang, Li Huang +1

Southeast University, Nanjing, China · Peking University, Beijing, China · Chongqing University, Chongqing, China

Language Model Steering LLM Safety Activation Steering

Sep 28, 2026cs.AI

AdaGuard: An Adaptive Guard Model with User-defined Policies

Guard models support the safe deployment of language model agents, but fixed risk taxonomies limit their ability to accommodate requirements that vary across applications and tasks. Under user-defined policies, detecting violations requires interpreting both the applicable rules and the agent's behavior, since identical actions can receive different judgments under different policies. To support learning this capability, we introduce AdaptiveSafety, a dataset of 10,939 training examples and 1,000 test examples covering policies with 1--100 rules. The dataset combines trajectories from multiple sources with policy and behavioral counterfactuals, pairing each example with an explanation and the complete set of violated rules. These counterfactuals expose changes that alter compliance, while structural augmentations provide supervision for consistency under rule reordering and identifier remapping. Building on this supervision, we propose SafePO, a reinforcement learning algorithm for refining violation identification while balancing explanatory reasoning and final verdicts. SafePO uses structured rewards to assess prediction correctness, retains group-relative advantages at the response level, and employs a separately trained value model to modulate token weights within explanation and verdict regions. Separate normalization controls their relative contribution to training despite differences in length. Through supervised initialization followed by SafePO, we develop AdaGuard, a family of 0.6B, 4B, and 8B guard models that assess agent trajectories under policies supplied at inference time. Our 4B model achieves binary accuracies of 89.30% on AdaptiveSafety and 71.82% on DynaBench. The project repository is available at https://github.com/Yunhao-Feng/AdaGuard

Yunhao Feng, Yifan Ding, Yuxiang Xie +4

National University of Defense Technology · Fudan University

LLM Guardrails AI Agent Safety Safety Filtering LLM Safety AI Agent Safety Benchmarks

Sep 27, 2026cs.LG

Quantifying Behavioral Tails in Black-Box Language Models

We introduce RareTrap, a framework for estimating the probability of severe behaviors in black box large language models (LLMs). A key challenge for probability estimation is defining a tractable distribution over the input space. To accomplish that, RareTrap uses a surrogate LLM and constructs a geometry-aware mapping from a lower-dimensional latent reference space into its token-embedding space to induce an explicit and reproducible distribution over input prompts. A response-level performance function is utilized on the response to quantify behavior severity. This enables sequential rare event simulation that concentrates evaluations on progressively more severe behaviors while preserving probability under the induced prompt distribution, which would otherwise be prohibitive to measure. Across 10 open-weight and two frontier models (GPT-5.4 and Claude Sonnet 4.6), we find that RareTrap successfully induces severe resource consumption behaviors and computes their probability with as few as 200 evaluations. RareTrap provides model developers a principled approach for evaluating language models under a common distribution, and prioritizing alignment effort to improve safety and mitigate risks.

Elsayed Eshra, Ali Al-Lawati, Dongwon Lee +1

The Pennsylvania State University University Park, PA 16802, USA

Language Model Safety Evaluation LLM Safety Rare-Event Estimation

Sep 24, 2026cs.LG

Beyond Average Safety: Chance-Constrained LLM Fine-tuning

Fine-tuning large language models on new objectives can improve helpfulness, instruction following, or domain-specific performance, but it can also induce regressions on safety-critical prompts. Existing safety-preserving fine-tuning methods typically control average safety loss or use weighted auxiliary penalties, which can obscure rare but severe failures. We propose a chance-constrained formulation for safety-preserving fine-tuning that limits the fraction of safety examples whose degradation relative to a reference model exceeds a prescribed threshold. Because the resulting empirical chance constraint contains a discontinuous indicator, we introduce a differentiable majorization of the violation rate, yielding a tractable conservative constraint. We then develop a constraint-aware gradient descent method that treats the majorized constraint as a safe set in parameter space and minimally modifies the fine-tuning direction to preserve feasibility. The resulting update admits a closed form and produces a tail-aware safety correction that emphasizes examples near or above the degradation threshold. We conduct an extensive set of experiments on harmful fine-tuning across three different tasks and three models and show that our approach consistently outperforms the baselines that exist in the literature. These results suggest that safety preservation in LLM fine-tuning is better viewed as a reliability-constrained optimization problem than as average-risk regularization.

Taha Entesari, Mahyar Fazlyab

Department of Electrical and Computer Engineering Johns Hopkins University

Stochastic Optimization Chance-Constrained Optimization LLM Fine-Tuning LLM Reliability LLM Safety

Sep 21, 2026cs.LG

Rare Event Estimation via Iterative Unalignment

As agents are deployed with increased autonomy, even extremely rare events along their stochastic output trajectories can occur and prove catastrophic. Safe deployment therefore does not depend on whether these events can occur, but on how often they might. We study the problem of estimating the probability of rare events that arise from stochastic variation in the agent's own actions. Estimating this type of risk requires searching over the combinatorially vast space of trajectories. Naive Monte Carlo is computationally prohibitive in this regime, and constructing effective importance sampling (IS) proposals requires coordinated changes to a context-dependent chain of conditional distributions. We develop a new IS method that perturbs the original model's weights to construct the proposal. The proposal is itself a differentiably parameterized language model, enabling gradient-based search over weight space. We formulate an objective that combines a differentiable surrogate for event amplification and an adaptive regularization scheme that dynamically balances amplification against estimator stability. We evaluate our approach on

\sim

120M and

\sim

2.6B models across three event families spanning 300+ rare events as rare as

10^{-9}

, with reference probabilities computed with

<10\%

relative standard error. In our most verifiable settings, we observe that our IS estimator achieves over

800\times

compute-weighted efficiency gains over naive Monte Carlo for events with probabilities lower than

10^{-7}

. Our implementation is available at https://github.com/namkoong-lab/iterative-unalignment.

Hanming Yang, Daksh Mittal, Jing Dong +1

Decision, Risk, and Operations Division, Columbia Business School

Importance Sampling LLM Safety Rare-Event Estimation

Sep 20, 2026cs.CL

Measuring the Assistant's Harmlessness Preferences on the User Turn

Post-training turns a general next-token predictor into a chat model with a persistent assistant persona. If that persona is a character the model plays only on its own turns, its preferences should govern what the assistant says, not what the model predicts other speakers will say. We test this boundary and find that it does not hold: a safety-relevant preference of the assistant---for harmless over harmful tasks---shapes the model's predictions even on the user's turn, where the assistant is not the one speaking. We find that this preference is small or near-zero in pretrained base models, that it emerges through post-training, replicated across open-weight model families, grows with scale, and can be moved by narrow finetuning that never touches user turns. We claim that this is evidence that post-training does not merely install a shallow assistant persona, but instead generalises beyond just the local assistant turn, into the model's representation of the user.

Jord Nguyen

Language Model Safety Evaluation LLM Safety LLM Post-Training

Sep 17, 2026cs.CL

Harm Laundering in GPT Models: Evidence That Gender Discrimination Is Transformed Rather Than Reduced Across Safety-Trained Generations

Safety evaluations for large language models rely on surface-form classifiers that report declining harm scores across model generations. We provide evidence that this methodology is systematically incomplete: explicit discriminatory content is transformed rather than removed. We call this \emph{harm laundering}. Analysing 450,000 gender-directed completions across 15 models spanning GPT-2 through to GPT-5 (OpenAI GPT lineage; three demographic conditions), we show that sexual violence clusters prevalent in GPT-2 women-directed output disappear by GPT-4, while men-directed completions gain positive representational territory (caregiving, emotional range, ally identity) that women-directed completions do not. The pattern is most visible at GPT-5: Topic~~5 (1,997~~documents) frames breast cancer as a men's rights debate, while zero equivalent clusters appear in women-directed output. Three independent classifiers score this content as non-toxic. Sentiment scores invert at GPT-4: early models demean women; later models over-correct. Topic diversity in women-directed completions falls 36% relative to men at the GPT-4 alignment boundary (W/M~

= 0.58

, from

0.91

at GPT-2). REGARD representational harm disparity correlates with release date (

ρ= +0.55

,

p = .034

) while Detoxify does not (

ρ= -0.23

,

p = .42

): toxicity scores fall as representational harm grows. We formalise harm laundering as a three-criteria test and provide a three-stage detection protocol applicable to any generative model. Within the OpenAI GPT lineage, toxicity score reduction is not a sufficient proxy for harm reduction.

Sarah Wyer, Sue Black, Noura Al Moubayed

Durham University

Gender Bias in Language Models Language Model Safety Evaluation LLM Safety Language Model Bias Evaluation

Sep 17, 2026cs.LG

Local Sparsity Enables Unsupervised LLM Safety Detection

Deployment-time safety methods for large language models (LLMs) are predominantly supervised and assume access to unsafe training data. Nevertheless, new attacks and harm categories regularly arise, not captured by models trained in such a supervised fashion. An alternative approach is to view this problem through the lens of anomaly detection, namely, to rely solely on modeling safe data and flagging out-of-distribution inputs. However, LLM activations lie in a high-dimensional space, raising concerns about whether anomaly detection is statistically feasible. We show that, under the linear representation hypothesis (LRH), there may indeed be hope. In the LRH concept space, which is typically recovered via a sparse autoencoder (SAE), nearby points share a small common active support. Using this local sparsity insight, we propose a framework for locally masked SAE-based anomaly detection, supported by theoretical justifications. We validate it on various architectures and datasets, including both capability-testing datasets and safety-specific datasets. Finally, when we allow algorithms to use 1% out-of-distribution data for calibration, locally sparse methods achieve near-optimal performance, demonstrating their ability to capture meaningful safety information while using only 1-2% of SAE neurons for computation.

Xin Chen, Gil Kur, Alexander Shevchenko +1

ETH Zürich

Sparse Autoencoders LLM Safety Unsupervised Anomaly Detection OOD Detection

Sep 16, 2026cs.AI

Safety Beyond the Interface: Detecting Harm via Latent States in Large Language Models

Autonomous systems increasingly rely on Large Language Models (LLMs) yet the safety infrastructure surrounding these models introduces latency and compute overhead. This limits utility in resource-constrained, time-critical deployments. Existing external guardrail models remain blind to the model's internal workings, creating a fundamental assurance gap. We ask: does the model already know when the content is harmful? We extract activations from LLaMA-3.1-8B and train lightweight MLP classifier probes (12.6M parameters) to detect harmful prompts. Evaluated on WildJailbreak, Beavertails, and AEGIS 2.0, our probes achieve F1 scores of 99%, 83%, and 84%, respectively competitive with 1000x larger guard models while cutting latency and compute costs.

Alizishaan Khatri, Chiquita Prabhu, Omkar Neogi

Wrynx Inc. · Independent Researcher

Safety Filtering LLM Safety Language Model Probing