Search: “safety”

Feature Geometry of LoRA Adapters: A Sparse Autoencoder Analysis of Representational Divergence in Fine-Tuned Language Models

arXiv:2605.28896v1 Announce Type: new Abstract: Low-Rank Adaptation (LoRA) has emerged as a widely adopted approach for adapting large language models, yet the internal representational changes induce

Content type

All Blog X Post Paper YouTube Reddit GitHub

Clear filters

Model Releases

FinGuard: Detecting Financial Regulatory Non-Compliance in LLM Interactions

DGX agent

arXiv:2605.29427v1 Announce Type: new Abstract: As large language models (LLMs) are increasingly deployed in financial services, a single non-compliant interaction can expose institutions to regulator

model-releasesarxiv-cs-cl

29 May 2026

GPIC: A Giant Permissive Image Corpus for Visual Generation

arXiv:2605.30341v1 Announce Type: cross Abstract: Studying scalable methods for visual generative modeling requires large, accessible, and stable datasets. We introduce GPIC, a Giant Permissive Image

How Reliable Are AI Attackers Against a Fixed Vulnerable Target? A 400-Run Empirical Study of LLM Penetration Testing Consistency

arXiv:2605.30096v1 Announce Type: cross Abstract: Large language models (LLMs) can autonomously conduct multi-stage cyber attacks, but the consistency of their offensive behavior under repeated trials

Latent Performance Profiling of Large Language Models

arXiv:2605.30018v1 Announce Type: new Abstract: Large language models (LLMs) frequently achieve impressive scores on standardized benchmarks, yet accuracy alone offers a limited view of their capabili

MechELK: A Mechanistic Interpretability Framework for Eliciting Latent Knowledge in Large Language Models

arXiv:2605.28825v1 Announce Type: new Abstract: Large language models (LLMs) frequently encode factual and reasoning knowledge in their internal representations that is not faithfully reflected in the

NICE: A Theory-Grounded Diagnostic Benchmark for Social Intelligence of LLMs

arXiv:2605.29685v1 Announce Type: new Abstract: As large language models (LLMs) are increasingly applied in social contexts such as emotional companionship and customer service, measuring their social

Recovering Diversity Without Losing Alignment: A DPO Recipe for Post-Trained LLMs

arXiv:2605.30021v1 Announce Type: new Abstract: Many open-ended instructions have multiple valid answers that users can benefit from seeing, but post-training often narrows an LLM's output space towar

Steering at the Source: Style Modulation Heads for Robust Persona Control

arXiv:2603.13249v2 Announce Type: replace-cross Abstract: Activation steering offers a computationally efficient mechanism for controlling Large Language Models (LLMs) without fine-tuning. While effec

SURGENT: A Surgical Multi-Agent Assistance System Across the Perioperative Workflow

arXiv:2605.29368v1 Announce Type: cross Abstract: The intricate nature of modern surgical care necessitates intelligent systems that can synthesize extensive patient records, support collaborative dec

TANDEM: Temporal-Aware Neural Detection for Multimodal Hate Speech

arXiv:2601.11178v2 Announce Type: replace Abstract: Social media platforms are increasingly dominated by long-form multimodal content, where harmful narratives are constructed through a complex interp

Debate with Images: Detecting Deceptive Behaviors in Multimodal Large Language Models

arXiv:2512.00349v2 Announce Type: replace Abstract: Are frontier AI systems becoming more capable? Certainly. Yet such progress is not an unalloyed blessing but rather a Trojan horse: behind their per

Localizing Input Uncertainty Quantification for Large Language Models via Shapley Values

arXiv:2605.28170v1 Announce Type: new Abstract: As large language models (LLMs) are increasingly integrated into high-stakes decision-making, the ability to reliably quantify uncertainty has become a

Multi-Adapter Representation Interventions via Energy Calibration

arXiv:2605.28722v1 Announce Type: new Abstract: Representation intervention has emerged as a promising paradigm for aligning large language models toward desired behaviors without modifying model weig

PAST2HARM: A Simple Adaptive Past Tense Attack for Jailbreaking Multimodal AI

arXiv:2605.27545v1 Announce Type: new Abstract: Jailbreak attacks on multimodal AI systems remain underexplored, even though unsafe image generation can have more severe consequences than unsafe text

Persuade Me if You Can: A Framework for Evaluating Persuasion Effectiveness and Susceptibility Among Large Language Models

arXiv:2503.01829v4 Announce Type: replace-cross Abstract: Large Language Models (LLMs) demonstrate persuasive capabilities that rival human-level persuasion. While these capabilities can be used for s

Edge AI Deployment Beyond Models: A BSP-Aware Systems Framework for Industrial Embedded Platforms

arXiv:2605.26119v1 Announce Type: cross Abstract: Industrial Edge AI programs often begin with the model and only later confront the platform. That sequencing is attractive because it allows early dem

On the Sensitivity of Instruction-tuned LLMs to Harmful Sentences in Long Inputs

arXiv:2510.05864v2 Announce Type: replace Abstract: Large language models (LLMs) increasingly operate on long inputs, yet their behavior when harmful sentences are sparsely embedded within such inputs

Persistent AI Agents in Academic Research: A Single-Investigator Implementation Case Study

arXiv:2605.26870v1 Announce Type: cross Abstract: Background: Large language models are typically evaluated as models, benchmarks, or short conversational episodes. Less is known about what happens wh

SEC-bench Pro: Can Language Models Solve Long-Horizon Software Security Tasks?

arXiv:2605.26548v1 Announce Type: cross Abstract: Large language models (LLMs) now support automated software security tasks, including vulnerability discovery and proof-of-concept (PoC) generation. E

Sentinel: Embodied Cooperative Spatial Reasoning and Planning

arXiv:2605.26239v1 Announce Type: new Abstract: In this work, we study Cooperative Spatial Intelligence, the ability of decentralized embodied agents to coordinate effectively under dynamic environmen

What Molecular Structure Cannot Tell Us: A Taxonomy of Explainability Gaps in GNN-Based Drug Toxicity Prediction

arXiv:2605.26183v1 Announce Type: cross Abstract: Graph Neural Networks (GNNs) have emerged as a structurally natural approach for molecular toxicity prediction, operating directly on atomic connectiv

AI Content Moderation in Therapy Conversations

arXiv:2605.25454v1 Announce Type: cross Abstract: Large language models (LLMs) are increasingly being used for emotional support. They are also being developed for formal therapy purposes. However, LL

Benchmarking and Learning Real-World Customer Service Dialogue

arXiv:2510.22143v3 Announce Type: replace Abstract: Existing benchmarks and training pipelines for industrial intelligent customer service (ICS) remain misaligned with real-world dialogue requirements

Emotional intelligence in large language models is fragmented across perception, cognition, and interaction

arXiv:2605.24686v1 Announce Type: new Abstract: As large language models (LLMs) are increasingly integrated into emotionally sensitive domains, the structural integrity of their emotional intelligence

MDIA: A Multi-Agent Diagnostic Intelligence Pipeline on HealthBench Professional

arXiv:2605.24699v1 Announce Type: new Abstract: Most reported gains on agentic-LLM clinical benchmarks are often attributed to prompt engineering, yet our results suggest that larger improvements can

PEDESTRIANQA: A Benchmark for Vision-Language Models on Pedestrian Intention and Trajectory Prediction

arXiv:2605.24562v1 Announce Type: cross Abstract: Pedestrian intention and trajectory prediction are critical for the safe deployment of autonomous driving systems, directly influencing navigation dec

Reward-free Alignment for Conflicting Objectives

arXiv:2602.02495v3 Announce Type: replace-cross Abstract: Direct alignment methods are increasingly used to align large language models (LLMs) with human preferences. However, many real-world alignmen

Security in the Fine-Tuning Lifecycle of Large Language Models: Threats, Defenses,Evaluation, and Future Directions

arXiv:2605.25073v1 Announce Type: cross Abstract: Background: Fine-tuning is central to adapting pre-trained Large Language Models (LLMs) to downstream tasks, but its reliance on training data, parame

SemanticZip: A Pilot Framework for Lossy Text Compression with LLMs as Semantic Decompressors

arXiv:2605.24541v1 Announce Type: cross Abstract: Text compression for large language model (LLM) systems is usually framed as token deletion, retrieval, summarization, or exact reconstruction. We stu

Understanding and Mitigating Premature Confidence for Better LLM Reasoning

arXiv:2605.24396v1 Announce Type: new Abstract: Long chains of thought (CoT) from current language models frequently contain logical gaps and unjustified leaps, limiting the gains from additional test

Cultural Adaptation in Large Language Models for Political Discourse

arXiv:2605.23332v1 Announce Type: new Abstract: The integration of large language models into political discourse analysis creates new opportunities for comparative research, policy analysis, and civi

DART: Semantic Recoverability for Structured Tool Agents

arXiv:2605.23311v1 Announce Type: new Abstract: When a structured tool agent fails mid-execution, the runtime faces a dilemma: replaying the entire task is safe but wasteful, while restoring from a lo

The Misattribution Gap: When Memory Poisoning Looks Like Model Failure in Agentic AI Systems

arXiv:2605.22842v1 Announce Type: cross Abstract: Multi-agent AI pipelines typically assume that agent misconduct originates from model misalignment. We identify a structural failure in this assumptio

Using Ensemble Diffusion to Estimate Uncertainty for End-to-End Autonomous Driving

arXiv:2506.00560v2 Announce Type: replace-cross Abstract: End-to-end planning systems for autonomous driving are rapidly improving, especially in closed-loop simulation environments like CARLA. Many s

MoralityGym: A Benchmark for Evaluating Hierarchical Moral Alignment in Sequential Decision-Making Agents

arXiv:2602.13372v2 Announce Type: replace-cross Abstract: Evaluating moral alignment in agents navigating conflicting, hierarchically structured human norms is a critical challenge at the intersection

23 May 2026

Beyond Benchmark Islands: Toward Representative Trustworthiness Evaluation for Agentic AI

arXiv:2603.14987v2 Announce Type: replace Abstract: Agentic AI systems increasingly act through tool-augmented, multi-step workflows whose failures (unsafe tool use, unauthorised actions, social harm)

Check Your LLM's Secret Dictionary! Five Lines of Code Reveal What Your LLM Learned (Including What It Shouldn't Have)

arXiv:2605.22005v1 Announce Type: cross Abstract: We show that singular value decomposition of the lm_head} weight matrix of a transformer-based large language model -- requiring only five lines of Py

MAVEN: A Multi-stage Agentic Annotation Pipeline for Video Reasoning Tasks

arXiv:2605.21917v1 Announce Type: new Abstract: Training Vision Language Models (VLMs) for video event reasoning requires high-quality structured annotations capturing not only what happened, but when

MOTOR: A Multimodal Dataset for Two-Wheeler Rider Behavior Understanding

arXiv:2605.22550v1 Announce Type: new Abstract: Two-wheelers account for a disproportionately high share of road fatalities in the Global South. Research on two-wheeler rider behavior, however, lags f

Pre-VLA: Preemptive Runtime Verification for Reliable Vision-Language-Action and World-Model Rollouts

arXiv:2605.22446v1 Announce Type: new Abstract: While large vision-language-action (VLA) models and generative world models (WM) have advanced long-horizon embodied intelligence, their practical deplo

Seizure-Semiology-Suite (S3): A Clinically Multimodal Dataset, Benchmark, and Models for Seizure Semiology Understanding

arXiv:2605.21852v1 Announce Type: new Abstract: While Multimodal Large Language Models (MLLMs) have demonstrated remarkable proficiency in general video understanding, their capacity to interpret invo

AgentAtlas: Beyond Outcome Leaderboards for LLM Agents

arXiv:2605.20530v1 Announce Type: cross Abstract: Large language model agents now act on codebases, browsers, operating systems, calendars, files, and tool ecosystems, but the benchmarks used to evalu

Agentic Physical AI toward a Domain-Specific Foundation Model for Nuclear Reactor Control

arXiv:2512.23292v3 Announce Type: replace-cross Abstract: The prevailing paradigm in AI for physical systems (scaling general-purpose foundation models toward universal multimodal reasoning) confronts

Conformal Selective Acting: Anytime-Valid Risk Control for RLVR-Trained LLMs

arXiv:2605.20270v1 Announce Type: new Abstract: A local specialist LLM, fine-tuned with reinforcement learning from verifiable rewards (RLVR) on operator-local data, is installed in a regulated organi

local-aiarxiv-cs-lg

LoCar: Localization-Aware Evaluation of In-Vehicle Assistants through Fine-Grained Sociolinguistic Control

arXiv:2605.21086v1 Announce Type: new Abstract: While Large Language Models (LLMs) are increasingly integrated into in-vehicle conversational systems, identifying the optimal model remains challenging

local-aiarxiv-cs-cl

MONET: A Massive, Open, Non-redundant and Enriched Text-to-image dataset

arXiv:2605.21272v1 Announce Type: new Abstract: Training large text-to-image models requires high-quality, curated datasets with diverse content and detailed captions. Yet the cost and complexity of c

Toxic Subword Pruning for Dialogue Response Generation on Large Language Models

arXiv:2410.04155v2 Announce Type: replace Abstract: How to defend large language models (LLMs) from generating toxic content is an important research area. Yet, most research focused on various model