The paper introduces ToxiAlert-Bench, a large-scale audio dataset that uniquely annotates both textual and paralinguistic sources of toxicity, and proposes a dual-head neural network that significantl…

The paper demonstrates that increasing the toxicity of prompts significantly degrades the factual reliability of LLMs, a degradation linked to the selective amplification of perturbation-sensitive nod…

The paper introduces retraining-free frameworks (Meow2X and TRNE) that mechanistically localize and suppress toxicity within language models by analyzing activation differences, achieving safety impro…

This study evaluated Roblox's chat moderation system using a large corpus of 2 million messages, finding that numerous unsafe messages related to grooming, harassment, and self-harm continue to escape…

The paper analyzes Codes of Conduct (CoCs) for online video games using a novel pipeline, finding that most multiplayer games lack CoCs despite safety needs, and that CoCs often lack specificity regar…

The paper introduces SuiChat-CN, a novel Chinese group-chat benchmark for contextual suicide risk assessment, demonstrating that multi-party conversational context is crucial for accurate detection.

This paper systematically evaluates how LLMs uncritically adapt to potentially dangerous user prompts related to eating disorders, finding that specific linguistic cues significantly increase the like…

STARE introduces a novel hierarchical reinforcement learning framework that treats the entire image generation process (denoising trajectory) as an attack surface, significantly improving the detectio…

The paper proposes a unified, microservices-based platform that integrates endpoint telemetry and predictive NLP models to provide real-time, correlated alerting for security risks and hate speech.

BraveGuard is a self-evolving defense framework that significantly improves the safety monitoring of computer-use agents by generating guard model supervision from open-world threat discovery and real…

BraveGuard is a self-evolving defense framework that improves the safety of computer-use agents by training guard models on open-world, multi-step threat trajectories rather than static benchmarks.

This paper introduces HarmAmp, a new benchmark for multi-turn harm amplification, and proposes TrajSafe, a proactive monitoring system that significantly reduces harmfulness in LLM interactions while…

This paper presents HarmfulSkillBench, a large-scale benchmark demonstrating that even small percentages of publicly available skills can be misused for harmful actions, significantly lowering LLM ref…

The paper introduces EUDAIMONIA, a new framework and benchmark for evaluating how well LLMs align with user welfare in social interactions, finding that even state-of-the-art models frequently violate…

The paper introduces a validated, consensus-labeled prompt bank that separates requests for executable malicious code (weapons) from requests for general harmful security knowledge, providing a more g…

This paper introduces a novel, comprehensive dataset that logs various cheating activities, including difficult-to-detect network flow disruption cheats, for the purpose of developing robust detection…

The paper introduces MGTEVAL, a comprehensive and extensible platform designed to systematically evaluate the performance, robustness, and efficiency of machine-generated text detectors.

The paper introduces CyBiasBench, a comprehensive benchmark that quantifies the inherent, agent-specific bias in LLM agents' attack selection patterns in cybersecurity scenarios.

The paper introduces SkillHarm, a comprehensive benchmark and automated framework for evaluating skill-based attacks across the entire agent skill-use lifecycle, demonstrating that current agents rema…

The paper introduces SkillVetBench, a novel two-stage benchmark that effectively detects and verifies malicious behavior in open agentic skill ecosystems, significantly outperforming existing static a…