Papers

New on arXiv

Sun 15 Mar 2026

255 papers

SAGE: Sensor-Augmented Grounding Engine for LLM-Powered Sleep Care AgentLLM agents and tool use → Medical and scientific domain agents
Tool-MCoT: Tool Augmented Multimodal Chain-of-Thought for Content Safety ModerationLLM reasoning → Reasoning for agents, planning and code +1
Inclusion-of-Thoughts: Mitigating Preference Instability via Purifying the Decision SpaceLLM reasoning → Reasoning benchmarks and evaluation
Whether, Not Which: Mechanistic Interpretability Reveals Dissociable Affect Reception and Emotion Categorization in LLMsMechanistic interpretability → Activation steering methods
Efficient Embedding-based Synthetic Data Generation for Complex Reasoning TasksLLM reasoning → Reasoning benchmarks and evaluation +1
FactorSmith: Agentic Simulation Generation via Markov Decision Process Decomposition with Planner-Designer-Critic RefinementLLM reasoning → Reasoning for agents, planning and code +1
The Provenance Paradox in Multi-Agent LLM Routing: Delegation Contracts and Attested Identity in LDPLLM agents and tool use → Agent threats, guardrails and governance
DASH: Dynamic Audio-Driven Semantic Chunking for Efficient Omnimodal Token CompressionEfficient LLM inference and serving → Audio and omni-modal token compression
State-Dependent Safety Failures in Multi-Turn Language Model InteractionLLM alignment / RLHF / preference optimisation → Safety alignment and jailbreak attacks +1
Comparative Analysis of 3D Convolutional and 2.5D Slice-Conditioned U-Net Architectures for MRI Super-Resolution via Elucidated Diffusion ModelsDiffusion models → Medical imaging and segmentation
ViDscribe: Multimodal AI for Customizing Audio Description and Question Answering in Online VideosMultimodal LLMs / vision-language models → Video understanding with MLLMs
VisionCoach: Reinforcing Grounded Video Reasoning via Visual-Perception PromptingKnowledge distillation → Multimodal distillation
Spectrum Matching: a Unified Perspective for Superior Diffusability in Latent DiffusionDiffusion models → Diffusion foundations and sampling
Nudging Hidden States: Training-Free Model Steering for Chain-of-Thought Reasoning in Large Audio-Language ModelsLLM reasoning → Audio-language model reasoning
Compute Allocation for Reasoning-Intensive Retrieval AgentsLLM reasoning → Knowledge- and tool-augmented reasoning +1
When Scanners Lie: Evaluator Instability in LLM Red-TeamingLLM safety and security → Jailbreaking
s2n-bignum-bench: A practical benchmark for evaluating low-level code reasoning of LLMsLLM reasoning → Mathematical reasoning
Delightful Policy GradientDeep reinforcement learning methods → RL theory and sample complexity
Tactile Modality Fusion for Vision-Language-Action ModelsVision-language-action models → Tactile, force-aware and humanoid VLA
PA3: Policy-Aware Agent Alignment through Chain-of-ThoughtLLM reasoning → Reasoning for agents, planning and code +1
A Loss Landscape Visualization Framework for Interpreting Reinforcement Learning: An ADHDP Case StudyDeep reinforcement learning methods → RL theory and sample complexity
D-MEM: Dopamine-Gated Agentic Memory via Reward Prediction Error RoutingLLM agents and tool use → Agent memory systems
Adapting Critic Match Loss Landscape Visualization to Off-policy Reinforcement LearningDeep reinforcement learning methods → RL theory and sample complexity
SuperLocalMemory V3: Information-Geometric Foundations for Zero-LLM Enterprise Agent MemoryLLM agents and tool use → Agent memory systems
Machine Learning-Driven Intelligent Memory System Design: From On-Chip Caches to StorageDeep reinforcement learning methods → RL for energy, networks and operations
Medical Image Spatial Grounding with Semantic SamplingMultimodal LLMs / vision-language models → Medical vision-language models
MorFiC: Fixing Value Miscalibration for Zero-Shot Quadruped TransferDeep reinforcement learning methods → RL for robot locomotion and manipulation
ASAP: Attention-Shift-Aware Pruning for Efficient LVLM InferenceMultimodal LLMs / vision-language models → Visual token pruning and efficiency +1
Expert Mind: A Retrieval-Augmented Architecture for Expert Knowledge Preservation in the Energy SectorRetrieval-augmented generation → RAG pipeline design and evaluation
Distilling Latent Manifolds: Resolution Extrapolation by Variational AutoencodersKnowledge distillation → General knowledge distillation methods
Visualizing Critic Match Loss Landscapes for Interpretation of Online Reinforcement Learning Control AlgorithmsDeep reinforcement learning methods → RL theory and sample complexity
Emotional Cost Functions for AI Safety: Teaching Agents to Feel the Weight of Irreversible ConsequencesLLM agents and tool use → General LLM agent frameworks and benchmarks
LatSearch: Latent Reward-Guided Search for Faster Inference-Time Scaling in Video DiffusionLLM reasoning → Diffusion reasoning and test-time scaling +2
VLA-Thinker: Boosting Vision-Language-Action Models through Thinking-with-Image ReasoningLLM reasoning → Spatial and embodied reasoning +1
Trust-Region Noise Search for Black-Box Alignment of Diffusion and Flow ModelsLLM alignment / RLHF / preference optimisation → Diffusion and multimodal model alignment
WorldVLM: Combining World Model Forecasting and Vision-Language ReasoningWorld models → World models for autonomous driving +1
Fine-tuning MLLMs Without Forgetting Is Easier Than You ThinkLLM reasoning → Multimodal reasoning +1
Infinite Problem Generator: Verifiably Scaling Physics Reasoning Data with Agentic WorkflowsLLM reasoning → Spatial and embodied reasoning +1
Unlearning-based sliding window for continual learning under concept driftLLM safety and security → Machine unlearning
Physics-Informed Policy Optimization via Analytic Dynamics RegularizationDeep reinforcement learning methods → RL for robot locomotion and manipulation
LongVidSearch: An Agentic Benchmark for Multi-hop Evidence Retrieval Planning in Long VideosLLM agents and tool use → Multimodal and embodied agents
AgentProcessBench: Diagnosing Step-Level Process Quality in Tool-Using AgentsLLM agents and tool use → Tool learning and function calling
Distilling Reasoning Without Knowledge: A Framework for Reliable LLMsLLM reasoning → Knowledge- and tool-augmented reasoning +1
Zoom to Essence: Trainless GUI Grounding by Inferring upon Interface ElementsMultimodal LLMs / vision-language models → Embodied and 3D vision-language models
GenState-AI: State-Aware Dataset for Text-to-Video Retrieval on AI-Generated VideosLLM reasoning → Multimodal reasoning
Questionnaire Responses Do not Capture the Safety of AI AgentsLLM agents and tool use → General LLM agent frameworks and benchmarks
ES-Merging: Biological MLLM Merging via Embedding Space SignalsMultimodal LLMs / vision-language models → MLLM training and visual instruction tuning
eNavi: Event-based Imitation Policies for Low-Light Indoor Mobile Robot NavigationOffline RL and imitation learning from data → Language-conditioned policies and VLAs
WestWorld: A Knowledge-Encoded Scalable Trajectory World Model for Diverse Robotic SystemsWorld models → Embodied and robot world models +1
From log π to π: Taming Divergence in Soft Clipping via Bilateral Decoupled Decay of Probability Gradient WeightLLM reasoning → Reinforcement learning for reasoning
SPARQ: Spiking Early-Exit Neural Networks for Energy-Efficient Edge AIEfficient LLM inference and serving → Early exit and edge inference
The Pulse of Motion: Measuring Physical Frame Rate from Visual DynamicsWorld models → Latent predictive world models
Trust Over Fear: How Motivation Framing in System Prompts Affects AI Agent Debugging DepthLLM agents and tool use → Social simulation and multi-agent debate
OxyGen: Unified KV Cache Management for VLA Inference under Multi-Task ParallelismEfficient LLM inference and serving → Visual token pruning in multimodal models
HomeGuard: VLM-based Embodied Safeguard for Identifying Contextual Risk in Household TaskMultimodal LLMs / vision-language models → Embodied and 3D vision-language models
Representation Alignment for Just Image Transformers is not Easier than You ThinkDiffusion models → Diffusion foundations and sampling
Exposing Long-Tail Safety Failures in Large Language Models through Efficient Diverse Response SamplingLLM alignment / RLHF / preference optimisation → Safety alignment and jailbreak attacks +1
Deconfounded Lifelong Learning for Autonomous Driving via Dynamic Knowledge SpacesEnd-to-end autonomous driving → End-to-end system architecture
AgroOmni: A Large-Scale Multi-view Agricultural Dataset for Cross-Scale Multimodal ReasoningLLM reasoning → Clinical and biomedical applications
On the Nature of Attention Sink that Shapes Decoding Strategy in Omni-LLMsLLM reasoning → Test-time compute and efficient thinking
UAVBench and UAVIT-1M: Benchmarking and Enhancing MLLMs for Low-Altitude UAV Vision-Language UnderstandingMultimodal LLMs / vision-language models → Embodied and 3D vision-language models
Data-Driven Physics Embedded Dynamics with Predictive Control and Reinforcement Learning for QuadrupedsDeep reinforcement learning methods → RL for robot locomotion and manipulation
ECG-Reasoning-Benchmark: A Benchmark for Evaluating Clinical Reasoning Capabilities in ECG InterpretationLLM reasoning → Clinical and biomedical applications
How Do Medical MLLMs Fail? A Study on Visual Grounding in Medical ImagesMultimodal LLMs / vision-language models → Medical vision-language models
Early Failure Detection and Intervention in Video Diffusion ModelsDiffusion models → Video generation
SemantiCache: Efficient KV Cache Compression via Semantic Chunking and Clustered MergingEfficient LLM inference and serving → KV cache optimization
Seeking Physics in Diffusion NoiseDiffusion models → Video generation
AEX: Non-Intrusive Multi-Hop Attestation and Provenance for LLM APIsLLM agents and tool use → Software engineering coding agents
Activation Surgery: Jailbreaking White-box LLMs without Touching the PromptLLM safety and security → Jailbreaking
Controllable Accent Normalization via Discrete DiffusionDiffusion models → Audio, speech and music
DiFlowDubber: Discrete Flow Matching for Automated Video Dubbing via Cross-Modal Alignment and SynchronizationDiffusion models → Audio, speech and music
MedPriv-Bench: Benchmarking the Privacy-Utility Trade-off of Large Language Models in Medical Open-Ended Question AnsweringRetrieval-augmented generation → Medical and biomedical RAG
Mitigating Overthinking in Large Reasoning Language Models via Reasoning Path Deviation MonitoringLLM reasoning → Test-time compute and efficient thinking +1
Why Do LLM-based Web Agents Fail? A Hierarchical Planning PerspectiveLLM agents and tool use → Web, search, and research agents
GoldenStart: Q-Guided Priors and Entropy Control for Distilling Flow PoliciesDiffusion models → Diffusion and flow policies for robotics +1
CamLit: Unified Video Diffusion with Explicit Camera and Lighting ControlDiffusion models → Video generation
QiMeng-CodeV-SVA: Training Specialized LLMs for Hardware Assertion Generation via RTL-Grounded Bidirectional Data SynthesisSynthetic data and self-improvement for LLMs → Code and instruction data synthesis
AeroGen: Agentic Drone Autonomy through Single-Shot Structured Prompting & Drone SDKLLM agents and tool use → Multimodal and embodied agents
Agentic DAG-Orchestrated Planner Framework for Multi-Modal, Multi-Hop Question Answering in Hybrid Data LakesLLM agents and tool use → Web, search, and research agents
"I'm Not Reading All of That": Understanding Software Engineers' Level of Cognitive Engagement with Agentic Coding AssistantsLLM agents and tool use → Software engineering coding agents
Self-Indexing KVCache: Predicting Sparse Attention from Compressed KeysEfficient LLM inference and serving → KV cache optimization
Membership Inference for Contrastive Pre-training Models with Text-only PII QueriesLLM safety and security → Membership inference attacks
FIND: A Simple yet Effective Baseline for Diffusion-Generated Image DetectionDiffusion models → Image generation, text-to-image and editing
Safety-Potential Pruning for Enhancing Safety Prompts Against VLM Jailbreaking Without RetrainingMultimodal LLMs / vision-language models → MLLM jailbreaks and adversarial robustness +1
Rethinking Evaluation in Retrieval-Augmented Personalized Dialogue: A Cognitive and Linguistic PerspectiveLLM reasoning → Knowledge- and tool-augmented reasoning +1
Memory as Asset: From Agent-centric to Human-centric Memory ManagementLLM agents and tool use → Agent memory systems
ChArtist: Generating Pictorial Charts with Unified Spatial and Subject ControlDiffusion models → Image generation, text-to-image and editing
DualTSR: Unified Dual-Diffusion Transformer for Scene Text Image Super-ResolutionDiffusion models → Inverse problems and image restoration
On Globally Optimal Stochastic Policy Gradient Methods for Domain Randomized LQR SynthesisDeep reinforcement learning methods → RL for games and combinatorial optimization
Relationship-Aware Safety Unlearning for Multimodal LLMsMultimodal LLMs / vision-language models → MLLM jailbreaks and adversarial robustness +1
Deeper Thought, Weaker Aim: Understanding and Mitigating Perceptual Impairment during Reasoning in Multimodal Large Language ModelsLLM reasoning → Multimodal reasoning +1
To LLM, or Not to LLM: How Designers and Developers Navigate LLMs as Tools or Teammates
From Natural Language to PromQL: A Catalog-Driven Framework with Dynamic Temporal Resolution for Cloud-Native Observability
Drift and selection in LLM text ecosystems
The Human Condition as Reflected in Contemporary Large Language Models
SensorPersona: An LLM-Empowered System for Continual Persona Extraction from Longitudinal Mobile Sensor Streams
DeepOFW: Deep Learning-Driven OFDM-Flexible Waveform Modulation for Peak-to-Average Power Ratio Reduction
Ternary Gamma Semirings: From Neural Implementation to Categorical Foundations
A Novel Framework using Intuitionistic Fuzzy Logic with U-Net and U-Net++ Architecture: A case Study of MRI Bain Image Segmentation
The Voice Behind the Words: Quantifying Intersectional Bias in SpeechLLMs

All 255