Papers

New on arXiv

Fri 28 Aug 2026

382 papers

AgenticCADedit: A Stateful, Tool-Mediated Agentic Approach to Multimodal 3D CAD EditingLLM agents and tool use → Multimodal and embodied agents
An Exploratory Ablation of a Small MLA--SSM Hybrid Language ModelMixture-of-experts → MoE large language models
STRAND: Benchmarking and Improving Object-Centric Spatio-Temporal Monitoring in Video Large Language ModelsMultimodal LLMs / vision-language models → Video understanding with MLLMs
PROVE: Proof-guided Regime-aware Operator Verification for Hallucination Detection in Medical Visual Question AnsweringMultimodal LLMs / vision-language models → Hallucination mitigation in LVLMs
When Agents Act Unwatched: The Reduced-Supervision Paradox in Agentic AILLM agents and tool use → Agent threats, guardrails and governance
Correlation-Guided Flow Matching with Annealed Masking for Spatial Transcriptomics GenerationDiffusion models → Medical imaging and segmentation
A Survey on Adversarial Attacks and Defenses for Diffusion Models Across Multiple ModalitiesDiffusion models → Robustness, privacy and adversarial aspects
The Vocabulary Gap Is an Equity Gap: Register Mismatch in Retrieval Systems for Public-Benefits AccessRetrieval-augmented generation → RAG robustness and adaptive retrieval
Imagine Before Retrieval: Prospective Skill Retrieval for LLM AgentsLLM agents and tool use → Agent skill learning and evolution
SocialBuddy: Tailoring Search Agent for Social ScenariosLLM agents and tool use → Web, search, and research agents
The information geometry of product-reference discrete diffusion: Interaction growth complexity and optimal schedulingDiffusion models → Diffusion foundations and sampling
Oculi: A Conversational Agentic Platform for Automated Credit Risk AnalysisLLM agents and tool use → Financial trading and forecasting agents
Multimodal Deep Learning for Uncertainty-Aware Radiation Pneumonitis Risk PredictionDiffusion models → Medical imaging and segmentation
Membership is Ownership: A Robust Ownership Verification Framework for Diffusion ModelsDiffusion models → Robustness, privacy and adversarial aspects +1
SemKV: Semantic Mixed-Precision KV Cache Quantization Guided by the Quality Cliff for Long-Context LLM InferenceEfficient LLM inference and serving → KV cache quantization
Hybrid Offline-Online Multi-Agent Decision Transformers for Wireless Resource ManagementOffline RL and imitation learning from data → Offline multi-agent RL
HALO: A Physics-Aware LLM Agent Framework for Nanophotonic DesignLLM agents and tool use → Medical and scientific domain agents
No Detectable Change in Side-Level WER from Prompt-Level Context: A Preregistered Ablation on a Production Oral-History CorpusMultimodal LLMs / vision-language models → MLLM training and visual instruction tuning
The Halt Vector: Internalizing a Causal Steering Intervention for Efficient ReasoningLLM reasoning → Test-time compute and efficient thinking
A rigor-matched audit of periodic-step layer skipping for efficient llm inference: conflayers versus swift, with a supplemental analysis of trained routing alternativesEfficient LLM inference and serving → Speculative decoding
Enhancing SAE-based Steering via Neighbor Integrated Feature SelectionMechanistic interpretability → SAE features and their applications
The reach of a verification tool decides its value: A controlled study of verification surface, artifact quality, and cost in AI coding agentsLLM agents and tool use → Software engineering coding agents
Efficient Geothermal Well-Control Optimization via Diffusion-Surrogate Reinforcement LearningDeep reinforcement learning methods → RL for energy, networks and operations
ASTRA - Agentic System for Ticket Resolution and AnalysisLLM agents and tool use → Agent threats, guardrails and governance
ClearText-Video: A Large-Scale Text-Centric Video Dataset Bridging Video Restoration and Scene-Text EnhancementMultimodal LLMs / vision-language models → Video understanding with MLLMs
ERR+: Sequential Entropy Resolution for Efficient and Decisive LLM ReasoningLLM reasoning → Reinforcement learning for reasoning
Beyond Vector Search: Comparing Classical RAG with Hybrid GraphRAG for Climate Science Q&ARetrieval-augmented generation → Agentic RAG, agents and memory
Inter-3D VQA: A Roadside Multimodal Benchmark for 3D Spatiotemporally Grounded Visual Question AnsweringLLM reasoning → Chain-of-thought reasoning for driving +1
Generation of High-Level Concepts in 3D Scene Graphs via Autoregressive DiffusionLLM reasoning → Spatial and embodied reasoning
Pro-Router: Token-Aware Progressive Model Routing with Adaptive Edge-Cloud Collaboration for Efficient Multimodal LLM InferenceMultimodal LLMs / vision-language models → Visual token pruning and efficiency
RoboPhys-3D: A Comprehensive Embodied World Model Evaluation via 3D ReconstructionWorld models → Embodied and robot world models
Video Generative Models as Geometry LearnerDiffusion models → 3D, point cloud and scene generation
An Enclosed Mode Is a Gauge Choice: Topology Relative to Reach in Certified Code World ModelsWorld models → World models for language agents
Training Communication-Efficient Mixture-of-Experts Language Models with Layer Re-ConfigurationMixture-of-experts → MoE large language models
Recognition Without Enforcement: Configuration-Dependent Failures in LLM Agent Instruction Arbitration and External ControlLLM agents and tool use → Agent threats, guardrails and governance
On the Maintenance and Co-evolution of Agent Plugins: An Empirical Study of Claude Code Plugin MarketplacesLLM agents and tool use → Software engineering coding agents
Ladders in Chaos: When, How, (and Perhaps Why) Does Test-Time Scaling Improve LLM Machine TranslationLLM reasoning → Test-time compute and efficient thinking
AcrossVAM1.0: Particle World Modeling for Text-Assisted Robot Video PredictionWorld models → Embodied and robot world models
LLM-Based Agents for Software and Systems Security: Approaches, Applications, and AssessmentLLM agents and tool use → Agent threats, guardrails and governance
NL2AGBench: Benchmarking LLM Auto-Formalization for AlphaGeometryLLM reasoning → Mathematical reasoning
ContextPilot: Teaching Agents for Proactive Context Management via Fine-grained RLLLM agents and tool use → Agent memory systems
LayerRecall: A State-Conditioned Memory Router for Long-Horizon Consistency in Video GenerationDiffusion models → Video generation
Learning to Use Tools: Reinforcement Learning for Tool-Integrated Mathematical ReasoningLLM reasoning → Reasoning for agents, planning and code +1
Sliding-window beats linear attentionEfficient LLM inference and serving → KV cache optimization
Fidelity Is Not Enough: Dispatch-Level Instrumentation for Agentic Datasheet ExtractionLLM agents and tool use → Web, search, and research agents
LUCID: An Agentic AI Framework on Digital-Twin in the Loop for QoS-Guaranteeing Robotic ControlLLM agents and tool use → Agentic AI for wireless networks
Linear Temporal Logic Translation via Human-Inspired Self-Constrained Reasoning for Robot Task SpecificationLLM reasoning → Spatial and embodied reasoning
LongPIBench: A Long-Context Benchmark for Prompt InjectionLLM safety and security → Prompt injection and agent security
How Far Can 5,500 Hours of Driving Take You? A Scaling Law Analysis of Video Diffusion ModelsDiffusion models → Diffusion and flow policies for robotics
VERA-8B: Evidence-Grounded Audit Risk Reasoning from SEC FilingsLLM reasoning → Reasoning for agents, planning and code
RetailAgent: Structured Adverse Timing in Self-Conditioned Multimodal LLM Trading AgentsLLM agents and tool use → Financial trading and forecasting agents
CamoDocs: A Poisoning Attack Against Retrieval-Augmented Language Models Using Camouflaged DocumentsRetrieval-augmented generation → Privacy and security of RAG +1
Semantic Head Specialization Guides Hybrid ViT Attention for Multimodal LLMsMultimodal LLMs / vision-language models → MLLM training and visual instruction tuning
PersonaForge: Realistic Multi-Turn User Simulation for Agentic SystemsLLM agents and tool use → Social simulation and multi-agent debate
EvoUndo: Recoverability-Constrained Self-Evolution for LLM Agent HarnessesLLM agents and tool use → Software engineering coding agents
GRACE:Gradient-guided Coreset Selection for LLM UnlearningLLM safety and security → Machine unlearning
AIM: Anchor Identity Features, Then Match for Multimodal Large Language Model UnlearningMultimodal LLMs / vision-language models → MLLM jailbreaks and adversarial robustness +1
VISTA: Verifier-Informed Student-to-Teacher Adaptation for On-Policy Self-DistillationKnowledge distillation → On-policy and self-distillation for LLMs
FUSED: Forensic-Semantic Mixture-of-Experts for AI Inpainting Detection and LocalizationMixture-of-experts → Vision mixture-of-experts
A Probabilistic Interpretation of KV Cache EvictionEfficient LLM inference and serving → KV cache optimization
LoopArena: Benchmarking Models as Runtime Controllers for Loop EngineeringLLM agents and tool use → Software engineering coding agents
Spatial-Semantic Reasoning using Large Language Models for Efficient UAV Search OperationsLLM reasoning → Spatial and embodied reasoning
CoCoBench: A Cooperative Coordination Benchmark for Embodied Multi-Agent Task PlanningLLM agents and tool use → Multimodal and embodied agents
Finding Where the Buck Stops: An Automated Failure Attribution-Based Reflection Framework for Multi-Agent CollaborationLLM agents and tool use → Software engineering coding agents
Physics-Guided Flow Matching for CT Image ReconstructionDiffusion models → Medical imaging and segmentation
Regime-Aware Portfolio Management via Retrieval-Augmented LLM-Guided Expert SwitchingRetrieval-augmented generation → Legal and financial RAG
Synth-JDoc: Synthesizing a Japanese Document Image Dataset for OCR with Diverse Layouts and Embedded ImagesMultimodal LLMs / vision-language models → Document understanding and multimodal RAG
REINS: Refusal-Enhanced Inhibitory Steering with Sparse Autoencoder FeaturesLLM safety and security → Safety alignment, refusal and guardrails +1
Focus Where It Counts: A Salience-Driven Vision-Language Model for Low Vision AssistanceMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning
WALDO: One-Shot Exemplar-Conditioned Object Detection in Cluttered ScenesWorld models → World modeling from LiDAR generation +1
EXPOSE: Explainable and Domain-Robust Embeddings from Pathology Vision Foundation Models using Sparse AutoencodersMechanistic interpretability → Interpreting biomedical and science models
Benchmarking large language model agent societies against human behavioural distributionsLLM agents and tool use → Social simulation and multi-agent debate
HARTS: Efficient Agentic Reinforcement Learning for Hybrid-Attention Models over Arbitrary Rollout TreesLLM agents and tool use → Reinforcement learning for LLM agents +1
Dual-Stream Semantic Guidance with Prototype Anchor Calibration for Source-Fully-Free Adaptation of Vision-Language ModelsMultimodal LLMs / vision-language models → Prompt learning and adaptation of CLIP +1
The Shape of Power: A Multilingual Framework for Social Power Reasoning in DialoguesLLM reasoning → Reasoning benchmarks and evaluation
Conditional Diffusion Models for Energy-Efficient DrivingDiffusion models → Diffusion and flow policies for robotics
Token-Budget Distillation: Transferring Full-Token Semantics to Compressed Video Vision-Language ModelsMultimodal LLMs / vision-language models → Visual token pruning and efficiency +1
VICT: Verifier-Instrumented Credit Tracing for Long-Horizon LLM Agent Reinforcement LearningLLM agents and tool use → Reinforcement learning for LLM agents
DeicticVLA: Unifying Instruction Modes Based on Language and Deictic Gestures in a Single VLAVision-language-action models → Reasoning and chain-of-thought in VLA
Do Medical Vision Models Reason About Anatomy? Probing the Spatial Inductive Biases of Learned Visual RepresentationsLLM reasoning → Clinical and biomedical applications
Ada-TokenCom: Rate-Adaptive Token Communications via Large-Model-Driven Token Compression and GenerationEfficient LLM inference and serving → Speculative decoding
Learning to Allocate Incentives for Incentivized Advertising via Offline Model-Based Reinforcement LearningOffline RL and imitation learning from data → Decision transformers and sequence modeling
WeAgent-MMSearch: Native Text-Vision Interaction for Multimodal Search AgentsLLM agents and tool use → Multimodal and embodied agents
Dynamic Alignment Compensation for Hallucination Mitigation in Large Vision-Language ModelsMultimodal LLMs / vision-language models → Hallucination mitigation in LVLMs
Characterization of Request and Token Energy Costs for LLM Inference Workloads on GPU PlatformsMixture-of-experts → Efficient MoE training and inference systems
Twin Worlds: Equivariance-Based Abstention for Evidence-Grounded ReasoningLLM reasoning → Reasoning benchmarks and evaluation
Coverage, Not Credit: Failure-Credit Routing of Zeroth-Order Perturbation Budgets Does Not Improve On-Pool Sample Efficiency for LLM AgentsLLM agents and tool use → Software engineering coding agents
When Can Conditional Flow Matching Replace Pointwise Negative Log-Likelihood?Diffusion models → Flow matching, time series and science
Visual Token Coding for Video Multimodal Large Language ModelsMultimodal LLMs / vision-language models → Visual token pruning and efficiency +1
Should I Use This Synthetic Dataset for Training? How to Test with Minimal Real DataWorld models → Embodied and robot world models
Moirae: A Multimodal Agent Collaborative Framework for Dynamic Android Malware DetectionLLM agents and tool use → Agent threats, guardrails and governance
CAITLYN: Can LLM Agents Autonomously Synthesize Defenses against Emerging Injection Attacks?LLM agents and tool use → Agent threats, guardrails and governance +1
When Evidence Shapes Collaboration: Knowledge-Conditioned Topology Generation for Multi-Agent SystemsLLM agents and tool use → General LLM agent frameworks and benchmarks
Learning from Hard Prompts: Difficulty-aware Advantage Amplification in Dynamic SamplingLLM reasoning → Reinforcement learning for reasoning
PhyMamba: Physics-Modulated Mamba for Robust Battery Health PrognosticsState-space sequence models (S4, Mamba) → Time series forecasting with SSMs
openJiuwen: Beyond Static Harnesses for Long-Horizon Coding AgentsLLM agents and tool use → Software engineering coding agents
AERA: Adaptive Evidence Residual Allocation for Efficient Test-Time ReasoningLLM reasoning → Test-time compute and efficient thinking
SABER: Stability-Aware Early Exit for LLM Reasoning via Adversarial Branch ProbingLLM reasoning → Test-time compute and efficient thinking +1
When Teacher Guidance Misleads: Reward-Aligned On-Policy DistillationKnowledge distillation → On-policy and self-distillation for LLMs
The Illusion of textitWhat If: Evaluating the Breakdown of Counterfactual Reasoning in LLMsLLM reasoning → Causal reasoning in LLMs

All 382