Knowledge catalogue
arXiv:2609.09798v1 Announce Type: cross Abstract: Large language models (LLMs) have been ex- ploited to generate malware, but the effective- ness of guardrails for code generation secu- rity remains u
arXiv:2610.12813v1 Announce Type: new Abstract: LLM agents increasingly operate over sensitive resources, such as private documents and mutable system states, which can create serious safety and secur
arXiv:2607.17575v3 Announce Type: replace Abstract: We propose ARBITER, a novel LLM guardrail framework that introduces two key ideas: (i) dual-hypothesis reasoning, a reasoning method for LLM guardra
arXiv:2610.10742v1 Announce Type: cross Abstract: AI systems increasingly rely on Large Language Models (LLMs) as core reasoning engines, making them targets for prompt injection and jailbreaks. Guard
arXiv:2610.12292v1 Announce Type: new Abstract: A typed decision model reads a piece of text and returns a probability over caller-defined options, each with a short written definition, generating no
arXiv:2608.27580v2 Announce Type: replace Abstract: Safety guardrails serve as the last line of defense against harmful inputs and outputs of large language models (LLMs), yet they are trained and eva
arXiv:2608.28646v1 Announce Type: new Abstract: Large language models (LLMs) can generate plausible-sounding ETF portfolios while silently violating basic KYC-style constraints on risk, fees, and dive
arXiv:2605.30693v1 Announce Type: cross Abstract: Building robust safety guardrails is essential for deploying Large Language Models across diverse real-world applications. However, this goal remains
arXiv:2605.31073v1 Announce Type: new Abstract: Reasoning-based LLM guardrails improve safety moderation by generating explicit rationales before issuing final decisions. However, their rationales do
arXiv:2605.30924v1 Announce Type: new Abstract: MLLM-powered embodied agents deployed in real-world environments encounter physical hazards. However, existing approaches lack explicit mechanisms for i
arXiv:2607.13081v1 Announce Type: cross Abstract: We present nsfaguard, a guardrail framework for securing agentic AI systems against operational threats, such as prompt injection, sensitive informati
arXiv:2606.26686v1 Announce Type: new Abstract: In order to screen a prompt or a response, the recent guardrail methods generate a chain-of-thought (CoT) before they issue a verdict. This design follo
arXiv:2605.17329v1 Announce Type: cross Abstract: Guardrails are a critical safety layer for modern AI systems, but their operating regime is changing. As LLMs are deployed as customized assistants, s
arXiv:2605.17610v1 Announce Type: cross Abstract: The rapid growth of online video platforms and AI-generated content has made reliable video guardrails a key challenge for safety and real-world deplo
arXiv:2605.17034v1 Announce Type: cross Abstract: Standard PII filters often miss contextual data leakage in RAG systems, such as non-regulated attribute clusters that collectively identify individual
arXiv:2605.17173v1 Announce Type: cross Abstract: Large language models exhibit safety degradation in non-English languages. Standard evaluation relies on Jailbreak Success Rate (JSR), which confounds
arXiv:2604.15579v1 Announce Type: cross Abstract: AI agents that interact with their environments through tools enable powerful applications, but in high-stakes business settings, unintended actions c
arXiv:2607.05910v1 Announce Type: cross Abstract: Image guardrails are typically trained and evaluated under a fixed safety policy, implicitly treating safety as an intrinsic property of an image. Rea
arXiv:2606.05936v1 Announce Type: new Abstract: Modern language models rely on pretraining filters to remove undesirable content from training corpora and inference-time guardrails to suppress undesir
arXiv:2510.13727v2 Announce Type: replace Abstract: Generative AI systems are increasingly assisting and acting on behalf of end users in practical settings, from digital shopping assistants to next-g
arXiv:2510.03485v2 Announce Type: replace Abstract: Autonomous web agents are increasingly deployed for long-horizon tasks, yet their ability to adhere to real-world policies remains critically undere
arXiv:2605.19940v1 Announce Type: new Abstract: Foundation models are increasingly deployed in socially sensitive domains such as education, mental health, and caregiving, where failures are often cum
arXiv:2609.29995v1 Announce Type: cross Abstract: AI teaching assistants (AI TAs) backed by large language models (LLMs) and pedagogical guardrails are increasingly being integrated into programming c
arXiv:2608.26171v1 Announce Type: cross Abstract: Multi-stage LLM hiring pipelines (resume improvement, interview question generation, answer feedback) can fabricate credentials, inflate qualifiers, a
arXiv:2608.15673v1 Announce Type: cross Abstract: Large language model guardrails can be viewed as policy-consistency problems: a system must determine which policy-relevant facts hold in a prompt-res
arXiv:2506.00166v2 Announce Type: replace-cross Abstract: Existing paradigms for ensuring AI safety, such as guardrail models and alignment training, often compromise either inference efficiency or de
arXiv:2604.25203v1 Announce Type: new Abstract: Deploying guardrails for custom policies remains challenging, as generic safety models fail to capture task-specific requirements, while prompting LLMs
arXiv:2604.16542v1 Announce Type: cross Abstract: Safety guardrails have become an active area of research in AI safety, aimed at ensuring the appropriate behavior of large language models (LLMs). How
arXiv:2609.16461v1 Announce Type: cross Abstract: Agentic large language model (LLM) systems rely on long interaction histories to preserve instructions, tool states, intermediate decisions, and unres
arXiv:2608.07167v1 Announce Type: new Abstract: Giving an AI agent the ability to send emails, query databases, or execute commands is useful--until the agent is tricked into doing something it should
arXiv:2607.22926v1 Announce Type: new Abstract: High-impact generative AI makes catastrophic misuse a lifecycle-control problem, not merely a prompt-filtering problem. SAGE is a safety-first, authoriz
arXiv:2605.29068v1 Announce Type: new Abstract: Maintaining the safety of large language models (LLMs) is crucial as they are increasingly deployed in real-world applications. Existing safety guardrai
arXiv:2604.07223v1 Announce Type: cross Abstract: As large language models (LLMs) evolve from static chatbots into autonomous agents, the primary vulnerability surface shifts from final outputs to int
arXiv:2609.01519v1 Announce Type: new Abstract: Interactive simulations increasingly evaluate policies in markets populated by language-model agents. Their outputs can look economic---prices, profits,
arXiv:2608.08485v1 Announce Type: new Abstract: Current LLM safety guardrails face a fundamental tension: fine-tuning distorts pre-trained representations while generative judges incur prohibitive inf
arXiv:2607.19361v1 Announce Type: cross Abstract: Most safety guardrails for large language models (LLMs) evaluate each prompt-response pair in isolation, which misses failures that arise only over a
arXiv:2607.19704v1 Announce Type: new Abstract: Scaling LLM-based applications to millions of users is bottlenecked by the inference cost and latency of modern foundation models. A natural fix is to c
arXiv:2607.19449v1 Announce Type: cross Abstract: Evaluation frameworks for tool-augmented LLM agents focus overwhelmingly on capability metrics or explicit tool crashes, leaving silent infrastructure
arXiv:2606.06837v1 Announce Type: cross Abstract: Scripted vs spontaneous speech detection is appealing for interview guardrails, but benchmark performance can be inflated by shortcuts tied to corpus
arXiv:2604.27275v1 Announce Type: cross Abstract: Large language model (LLM) reading assistants are increasingly used in settings that require interpretation rather than simple retrieval. In these con
arXiv:2609.33671v1 Announce Type: cross Abstract: When must a foundation-model safety gateway generate tokens, and when should it directly output a calibrated decision? We study calibrated standalone
arXiv:2609.24801v1 Announce Type: cross Abstract: Large language models (LLMs) are increasingly deployed in production systems, raising concerns about their exposure to adversarial manipulation throug
arXiv:2608.10279v1 Announce Type: cross Abstract: Streaming language-model output creates a release-timing problem: complete-response moderation acts after streamed text has escaped, whereas repeated
arXiv:2607.21401v1 Announce Type: cross Abstract: A vision-language AI assistant returns its answer as a stream of generated tokens. Therefore, a safety guard that watches that answer has to keep up w
arXiv:2607.02582v1 Announce Type: new Abstract: Generative image models are capable of producing images that bear a strong resemblance to, or replicate, recognizable intellectual property (IP). In thi
arXiv:2606.28360v1 Announce Type: cross Abstract: University students often struggle to navigate complex academic policies, leading to advising bottlenecks and delayed access to critical information.
arXiv:2606.02041v1 Announce Type: new Abstract: Large language models increasingly stream long, reasoning-intensive responses in real time, making when to moderate as critical as whether to moderate.
arXiv:2605.10633v1 Announce Type: cross Abstract: Fine-tuning Large Language Models (LLMs) on benign narrow data can sometimes induce broad harmful behaviors, a vulnerability termed emergent misalignm
arXiv:2610.09906v1 Announce Type: cross Abstract: Security Operations Centers (SOCs) for information technology and operational technology share one incident-response problem: a flood of correlated al
arXiv:2609.05535v1 Announce Type: new Abstract: Verdict-only evaluation does not reveal whether a vision-language model (VLM) used the visual evidence that should support its decision. We study this p
arXiv:2504.08192v2 Announce Type: replace Abstract: Machine unlearning is a promising approach to improve LLM safety by removing unwanted knowledge from the model. However, prevailing gradient-based u
arXiv:2609.02246v1 Announce Type: new Abstract: Self-improving agent pipelines have a problem at their center. An optimizer rewrites prompts to score higher, and the score comes from a judge that is i
arXiv:2608.24777v1 Announce Type: new Abstract: LLM-based agents can interact with external environments through tool invocation, but this capability also introduces security risks such as file modifi
arXiv:2605.22984v1 Announce Type: cross Abstract: Test-Time Training (TTT) is an emerging paradigm that enables models to adapt their parameters during inference, improving performance on tasks such a
arXiv:2605.21609v1 Announce Type: new Abstract: Large language models (LLMs) are increasingly embedded in adolescent digital environments, mediating information seeking, advice, and emotionally sensit
arXiv:2604.20436v1 Announce Type: cross Abstract: Generative AI (GenAI) is reshaping software engineering by shifting development from manual coding toward agent-driven implementation. While vibe codi
arXiv:2609.22048v1 Announce Type: new Abstract: A selective predictor acts as a safety gate: it returns an output only when the prediction appears sufficiently trustworthy. Deployments increasingly re
arXiv:2608.21880v1 Announce Type: new Abstract: Bangla large language model (LLM) safety is difficult to evaluate with English-centric or standard-script benchmarks because Bangla users routinely writ
arXiv:2608.06110v1 Announce Type: new Abstract: This paper presents ECHO (Enhanced Care & Health Observer), a locally-deployable conversational health assistant for long-term chronic care management.
arXiv:2601.14660v2 Announce Type: replace-cross Abstract: Agentic Large Language Models (LLMs) are models able to reason, plan, and execute tools over unstructured data. These abilities are enabling t