Papers

New on arXiv

Sat 28 Feb 2026

238 papers

MobiFlow: Real-World Mobile Agent Benchmarking through Trajectory FusionLLM agents and tool use → GUI and mobile agents
How AI Systems Think About Education: Analyzing Latent Preference Patterns in Large Language ModelsLLM reasoning → Reasoning benchmarks and evaluation +1
From Flat to Structural: Enhancing Automated Short Answer Grading with GraphRAGRetrieval-augmented generation → Knowledge-graph and structured-knowledge RAG
CURE: A Multimodal Benchmark for Clinical Understanding and Retrieval EvaluationMultimodal LLMs / vision-language models → Medical vision-language models
Pragma-VL: Towards a Pragmatic Arbitration of Safety and Helpfulness in MLLMsMultimodal LLMs / vision-language models → MLLM jailbreaks and adversarial robustness +1
RelayCaching: Accelerating LLM Collaboration via Decoding KV Cache ReuseLLM agents and tool use → Software engineering coding agents +1
From Stochastic Answers to Verifiable Reasoning: Interpretable Decision-Making with LLM-Generated CodeLLM reasoning → Reasoning benchmarks and evaluation
PaLMR: Towards Faithful Visual Reasoning via Multimodal Process AlignmentLLM reasoning → Multimodal reasoning +1
Roots Beneath the Cut: Uncovering the Risk of Concept Revival in Pruning-Based Unlearning for Diffusion ModelsDiffusion models → Robustness, privacy and adversarial aspects +1
When Scaling Fails: Mitigating Audio Perception Decay of LALMs via Multi-Step Perception-Aware ReasoningLLM reasoning → Audio-language model reasoning
Social-JEPA: Emergent Geometric IsomorphismWorld models → Latent predictive world models
Silent Sabotage During Fine-Tuning: Few-Shot Rationale Poisoning of Compact Medical LLMsLLM safety and security → Backdoor and data poisoning attacks
Constitutional Black-Box Monitoring for Scheming in LLM AgentsLLM agents and tool use → Agent threats, guardrails and governance
MME: Mixture of Mesh Experts with Random Walk Transformer GatingMixture-of-experts → Vision mixture-of-experts
COMBAT: Conditional World Models for Behavioral Agent TrainingWorld models → Latent dynamics models and model-based RL +1
A Comprehensive Evaluation of LLM Unlearning Robustness under Multi-Turn InteractionLLM safety and security → Machine unlearning
ContextCov: Deriving and Enforcing Executable Constraints from Agent Instruction FilesLLM agents and tool use → Software engineering coding agents
Curation Leaks: Membership Inference Attacks against Data Curation for Machine LearningLLM safety and security → Membership inference attacks
MetaMind: General and Cognitive World Models in Multi-Agent Systems by Meta-Theory of MindWorld models → World models for language agents
NERFIFY: A Multi-Agent Framework for Turning NeRF Papers into CodeLLM agents and tool use → Web, search, and research agents
Structure Matters: Evaluating Multi-Agents Orchestration in Generative Therapeutic ChatbotsLLM agents and tool use → Medical and scientific domain agents
Generalizing Score-based generative models for Heavy-tailed DistributionsDiffusion models → Diffusion foundations and sampling
Analyzing and Improving Fast Sampling of Text-to-Image Diffusion ModelsDiffusion models → Diffusion foundations and sampling
Exploratory Randomization for Discrete-Time Risk-Sensitive Benchmarked Investment Management with Reinforcement LearningDeep reinforcement learning methods → Adversarial robustness and RL in finance
UniHM: Unified Dexterous Hand Manipulation with Vision Language ModelMultimodal LLMs / vision-language models → Embodied and 3D vision-language models
Qwen3-Coder-Next Technical ReportLLM agents and tool use → Software engineering coding agents
RLAR: An Agentic Reward System for Multi-task Reinforcement Learning on Large Language ModelsLLM agents and tool use → Reinforcement learning for LLM agents +1
MARS: Harmonizing Multimodal Convergence via Adaptive Rank SearchMultimodal LLMs / vision-language models → MLLM training and visual instruction tuning
Keyframe-Guided Structured Rewards for Reinforcement Learning in Long-Horizon Laboratory RoboticsDeep reinforcement learning methods → RL for robot locomotion and manipulation
SkillCraft: Can LLM Agents Learn to Use Tools Skillfully?LLM agents and tool use → Agent skill learning and evolution
Frozen Policy Iteration: Computationally Efficient RL under Linear Qπ Realizability for Deterministic DynamicsDeep reinforcement learning methods → RL theory and sample complexity
Towards Khmer Scene Document Layout DetectionMultimodal LLMs / vision-language models → Document understanding and multimodal RAG
DRIV-EX: Counterfactual Explanations for Driving LLMsLLM reasoning → Chain-of-thought reasoning for driving +1
RAVEL: Reasoning Agents for Validating and Evaluating LLM Text SynthesisLLM reasoning → Reasoning for agents, planning and code +1
Act Like a Pathologist: Tissue-Aware Whole Slide Image ReasoningLLM reasoning → Clinical and biomedical applications +1
InfoPO: Information-Driven Policy Optimization for User-Centric AgentsLLM agents and tool use → Reinforcement learning for LLM agents
Mema: Memory-Augmented Adapter for Enhanced Vision-Language UnderstandingMultimodal LLMs / vision-language models → MLLM training and visual instruction tuning
LiTS: A Modular Framework for LLM Tree SearchLLM reasoning → Test-time compute and efficient thinking
TraceSIR: A Multi-Agent Framework for Structured Analysis and Reporting of Agentic Execution TracesLLM agents and tool use → Software engineering coding agents
TGM-VLA: Task-Guided Mixup for Sampling-Efficient and Robust Robotic ManipulationOffline RL and imitation learning from data → Robot manipulation imitation learning
CMI-RewardBench: Evaluating Music Reward Models with Compositional Multimodal InstructionLLM alignment / RLHF / preference optimisation → Preference alignment for recommenders and RAG
IdGlow: Dynamic Identity Modulation for Multi-Subject GenerationDiffusion models → Image generation, text-to-image and editing
Theory of Code Space: Do Code Agents Understand Software Architecture?LLM agents and tool use → Software engineering coding agents
UNICBench: UNIfied Counting Benchmark for MLLMMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning
LangGap: Diagnosing and Closing the Language Gap in Vision-Language-Action ModelsVision-language-action models → VLA architectures and generalization
Unlearning Evaluation through Subset Statistical IndependenceLLM safety and security → Machine unlearning
Draft-Thinking: Learning Efficient Reasoning in Long Chain-of-Thought LLMsLLM reasoning → Prompting and chain-of-thought reasoning
Efficient Long-Sequence Diffusion Modeling for Symbolic Music GenerationDiffusion models → Audio, speech and music
SWE-Hub: A Unified Production System for Scalable, Executable Software Engineering TasksLLM agents and tool use → Software engineering coding agents
CoMoL: Efficient Mixture of LoRA Experts via Dynamic Core Space MergingMixture-of-experts → Mixture of LoRA experts
MIDAS: Multi-Image Dispersion and Semantic Reconstruction for Jailbreaking MLLMsMultimodal LLMs / vision-language models → MLLM jailbreaks and adversarial robustness +1
Advancing Multimodal Judge Models through a Capability-Oriented Benchmark and MCTS-Driven Data GenerationMultimodal LLMs / vision-language models → MLLM training and visual instruction tuning
LOGIGEN: Logic-Driven Generation of Verifiable Agentic TasksLLM agents and tool use → General LLM agent frameworks and benchmarks
RAFM: Retrieval-Augmented Flow Matching for Unpaired CBCT-to-CT TranslationDiffusion models → Medical imaging and segmentation
DenoiseFlow: Uncertainty-Aware Denoising for Reliable LLM Agentic WorkflowsLLM reasoning → Diffusion reasoning and test-time scaling +1
Bridge Matching Sampler: Scalable Sampling via Generalized Fixed-Point Diffusion MatchingDiffusion models → Flow matching, time series and science
TP-Spikformer: Token Pruned Spiking TransformerEfficient LLM inference and serving → Token pruning and merging for ViTs
Mesh-Pro: Asynchronous Advantage-guided Ranking Preference Optimization for Artist-style Quadrilateral Mesh GenerationLLM alignment / RLHF / preference optimisation → Diffusion and multimodal model alignment
CIRCUS: Circuit Consensus under Uncertainty via Stability EnsemblesMechanistic interpretability → Circuits and mechanistic analysis
SIAgent: Spatial Interaction Agent via LLM-powered Eye-Hand Motion Intent Understanding in VRLLM agents and tool use → Multimodal and embodied agents
Phys-Diff: A Physics-Inspired Latent Diffusion Model for Tropical Cyclone ForecastingDiffusion models → Flow matching, time series and science
Jano: Adaptive Diffusion Generation with Early-stage Convergence AwarenessDiffusion models → Diffusion foundations and sampling
ProtegoFed: Backdoor-Free Federated Instruction Tuning with Interspersed Poisoned DataLLM safety and security → Backdoor and data poisoning attacks
MLLM-4D: Towards Visual-based Spatial-Temporal IntelligenceLLM reasoning → Spatial and embodied reasoning
Wavelet-based Frame Selection by Detecting Semantic Boundary for Long Video UnderstandingMultimodal LLMs / vision-language models → Video understanding with MLLMs
Multimodal Adaptive Retrieval Augmented Generation through Internal Representation LearningRetrieval-augmented generation → Multimodal RAG
What Do Visual Tokens Really Encode? Uncovering Sparsity and Redundancy in Multimodal Large Language ModelsMultimodal LLMs / vision-language models → Visual token pruning and efficiency
M2: Dual-Memory Augmentation for Long-Horizon Web Agents via Trajectory Summarization and Insight RetrievalMultimodal LLMs / vision-language models → Embodied and 3D vision-language models +1
WirelessAgent++: Automated Agentic Workflow Design and Benchmarking for Wireless NetworksLLM agents and tool use → Agentic AI for wireless networks
Antibody: Strengthening Defense Against Harmful Fine-Tuning for Large Language Models via Attenuating Harmful Gradient InfluenceLLM safety and security → Safety alignment, refusal and guardrails
ArtiFixer: Enhancing and Extending 3D Reconstruction with Auto-Regressive Diffusion ModelsDiffusion models → 3D, point cloud and scene generation
LifeEval: A Multimodal Benchmark for Assistive AI in Egocentric Daily Life TasksMultimodal LLMs / vision-language models → Embodied and 3D vision-language models
RAISE: Requirement-Adaptive Evolutionary Refinement for Training-Free Text-to-Image AlignmentDiffusion models → Image generation, text-to-image and editing
TokenCom: Vision-Language Model for Multimodal and Multitask Token CommunicationsMultimodal LLMs / vision-language models → Visual token pruning and efficiency
U-VLM: Hierarchical Vision Language Modeling for Report GenerationMultimodal LLMs / vision-language models → Medical vision-language models
Atomicity for Agents: Exposing, Exploiting, and Mitigating TOCTOU Vulnerabilities in Browser-Use AgentsLLM agents and tool use → Agent threats, guardrails and governance
Wireless Power Control Based on Large Language ModelsLLM reasoning → Test-time compute and efficient thinking
From Goals to Aspects, Revisited: An NFR Pattern Language for Agentic AI SystemsLLM agents and tool use → General LLM agent frameworks and benchmarks
Cloud-OpsBench: A Reproducible Benchmark for Agentic Root Cause Analysis in Cloud SystemsLLM agents and tool use → Software engineering coding agents
DreamWorld: Unified World Modeling in Video GenerationWorld models → Video-generation world models and simulators
ReMoT: Reinforcement Learning with Motion Contrast TripletsDeep reinforcement learning methods → RL for robot locomotion and manipulation
Improved Adversarial Diffusion Compression for Real-World Video Super-ResolutionDiffusion models → Inverse problems and image restoration
Test-Driven Agentic Framework for Reliable Robot ControllerLLM agents and tool use → Multimodal and embodied agents
HydroShear: Hydroelastic Shear Simulation for Tactile Sim-to-Real Reinforcement LearningDeep reinforcement learning methods → RL for robot locomotion and manipulation
Mamba-CAD: State Space Model For 3D Computer-Aided Design Generative ModelingState-space sequence models (S4, Mamba) → Video, motion and generative Mamba models
Self-Correction Inside the Model: Leveraging Layer Attention to Mitigate Hallucinations in Large Vision Language ModelsMultimodal LLMs / vision-language models → Hallucination mitigation in LVLMs
ROKA: Robust Knowledge Unlearning against AdversariesLLM safety and security → Machine unlearning
Taxonomy-Aware Representation Alignment for Hierarchical Visual Recognition with Large Multimodal ModelsMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning
LLM-Bootstrapped Targeted Finding Guidance for Factual MLLM-based Medical Report GenerationMultimodal LLMs / vision-language models → Medical vision-language models +1
Weight Updates as Activation Shifts: A Principled Framework for SteeringMechanistic interpretability → Activation steering methods
PointAlign: Feature-Level Alignment Regularization for 3D Vision-Language ModelsMultimodal LLMs / vision-language models → Embodied and 3D vision-language models
SSR: Pushing the Limit of Spatial Intelligence with Structured Scene ReasoningLLM reasoning → Spatial and embodied reasoning +1
Hereditary Geometric Meta-RL: Nonlocal Generalization via Task SymmetriesDeep reinforcement learning methods → RL for robot locomotion and manipulation
A Consistency-Centric Approach to Set-Based Optimization with Multiple Models of Unranked Fidelity
DiagramBank: A Quality-Audited Dataset of Scientific Schematic Diagrams with Multi-Level Document Context
Leveraging Computerized Adaptive Testing for Cost-effective Evaluation of Large Language Models in Medical Benchmarking
Improving Automatic Summarization of Radiology Reports through Mid-Training of Large Language Models
A Novel end-to-end Digital Health System Using Deep Learning-based ECG Analysis
Rubric-Guided Fine-tuning of SpeechLLMs for Multi-Aspect, Multi-Rater L2 Reading-Speech Assessment
Multi-Agent Reinforcement Learning for Dynamic Pricing: Balancing Profitability,Stability and Fairness

All 238