Papers

New on arXiv

Fri 4 Sep 2026

401 papers

H2LooP Telecom Model v1: From Telecom Comprehension to Autonomous Issue and PR ResolutionLLM agents and tool use → Web, search, and research agents
Evaluation of MLLM-Agnostic Plug-and-Play Keyframe Selection Methods for Long Video UnderstandingMultimodal LLMs / vision-language models → Video understanding with MLLMs
Occamy-1.0: Open Pareto-frontier 35B Intelligence for Co-workLLM agents and tool use → General LLM agent frameworks and benchmarks
DI-Bench: Systematically Generating In-Domain Data Intelligence Benchmarks for Enterprise AgentsLLM agents and tool use → Web, search, and research agents
Inference-Time Graph Engineering for Multi-Agent LLM WorkflowsLLM agents and tool use → General LLM agent frameworks and benchmarks
RAPTOR: Role-Aware Private Training for Mixture-of-ExpertsMixture-of-experts → Expert routing strategies
Interface-Aware KV Cache Quantization for Dense On-Chip NVM in Long-Context LLM DecodingEfficient LLM inference and serving → KV cache quantization
RAGMark: A Comprehensive Framework for Benchmarking Retrieval-Augmented Generation SystemsRetrieval-augmented generation → RAG pipeline design and evaluation
From Monolithic Blending to Agentic Orchestration: Dynamic Response for Conversational Assistants at ScaleLLM agents and tool use → General LLM agent frameworks and benchmarks
CrisisKD: Five-Stage Knowledge Distillation for Aspect-Level Sentiment and Emotion Analysis in Crisis DiscourseKnowledge distillation → Multimodal distillation
Concord: A Video Relational Algebra for Cross-Modal Query OptimizationMultimodal LLMs / vision-language models → Video understanding with MLLMs
SimTIO: A Simulation-Grounded Multi-Agent LLM Framework for Compositional Traffic Intervention OptimizationLLM agents and tool use → Multimodal and embodied agents
Beyond Prompts: Measuring and Optimizing LLM Tool-Agent HarnessesLLM agents and tool use → Tool learning and function calling
CUSP: Decomposable Collective Uncertainty for Multi-Agent Multimodal ReasoningLLM reasoning → Multimodal reasoning +2
Intra-Prompt Parallel Decoding for Common-Context Question AnsweringEfficient LLM inference and serving → Prompt and context compression
Regret Dominates Surprise: Design-Time Requirements Engineering for Agentic-AI SafetyLLM agents and tool use → Agent threats, guardrails and governance
Who Maintains Agent Skills? A Longitudinal Study of Human-Governed, AI-Assisted Skill MaintenanceLLM agents and tool use → Agent skill learning and evolution
Endogenous Exploration in Reinforcement Learning with Intrinsic CuriosityDeep reinforcement learning methods → Deep RL algorithms and representations
Better Together: Complementary Query Rewriting Under a Strong RAG BaselineRetrieval-augmented generation → Generation and context integration
Unsupervised Transfer Clustering for Mitigating Cold Start in Active Prompt LearningMultimodal LLMs / vision-language models → Prompt learning and adaptation of CLIP
SceneMosaic: Efficient and Diverse Simulation-Ready Scene Generation via Hybrid Agentic Layout EvolutionLLM agents and tool use → Multimodal and embodied agents
GE-Act 2.0: Pretraining and Scaling a World-Action Model for Robotic ManipulationWorld models → Embodied and robot world models
Agents Trust Tools Too Much: Measuring Reliance on Unreliable ToolsLLM agents and tool use → Agent threats, guardrails and governance
An overview of 3D Vision-Language ModelsMultimodal LLMs / vision-language models → Embodied and 3D vision-language models
EnvCraft: Synthesizing Executable Environments in Agentic RL for Claw-like AgentLLM agents and tool use → Reinforcement learning for LLM agents
Capsule Lens: Locating and Tracking Concept Geometry in Model RepresentationsMechanistic interpretability → Sparse autoencoder training and evaluation
Grounded Skill Synthesis from Code at Scale for Agentic IntelligenceLLM agents and tool use → Agent skill learning and evolution
Information-Guided Safe Reinforcement Learning for Autonomous Gas Source Localization using sUASDeep reinforcement learning methods → RL for robot navigation and UAV control
WearableQA: A Benchmark for Health Reasoning over Real-World Wearable DataLLM reasoning → Clinical and biomedical applications
Diffusion TV: Experiencing Diffusion Models through Tangible, Embodied InteractionDiffusion models → Diffusion and flow policies for robotics
Same Trajectory, Contradictory Rewards (ROBORMBENCH): Paraphrase Fragility in Vision Language Reward ModelsMultimodal LLMs / vision-language models → Embodied and 3D vision-language models +1
Multi-Step Tool-Calling over Korean Open Public APIs: A Benchmark and a Data-Synthesis RecipeLLM agents and tool use → Tool learning and function calling
Think-Verify-Revise: Neuro-Symbolic Visual Reasoning with Vision-Language Models and Dynamic Logic Tensor NetworksLLM reasoning → Multimodal reasoning +1
Reflection-aware Generative Novel View SynthesisDiffusion models → 3D, point cloud and scene generation
CUA-Universe: A Scalable and Dynamic Environment for Hybrid GUI+CLI AgentsLLM agents and tool use → GUI and mobile agents
Towards Neuro-Symbolic Procedural Reasoning for Long-Horizon Vision-Language-Action ManipulationVision-language-action models → Reasoning and chain-of-thought in VLA
Distill Globally, Adapt Locally: Reasoning Distillation and Product-Type Test-Time Training for Scalable Trade-Up RecommendationLLM reasoning → LLM reasoning for recommendation +1
MEOX: Compact Multimodal Mixture-of-Experts for Earth ObservationMixture-of-experts → Vision mixture-of-experts
Does Your Agent's Memory Survive a Model Upgrade? A Controlled Study of Memory PortabilityRetrieval-augmented generation → Agentic RAG, agents and memory
Lightweight Vision Transformer Compression for On-Device Plant Disease Detection in Resource-Constrained Agricultural Field ConditionsKnowledge distillation → Detection and dense prediction distillation
Machine Unlearning as Private Retroactive AlgorithmsLLM safety and security → Machine unlearning
RoboSPA: Can VLA Models Go Beyond Simple Scenes and Short-Horizon Tasks?Vision-language-action models → Generalist VLA models and benchmarks
Optimal Rates for Agentic Networked Information AggregationLLM agents and tool use → General LLM agent frameworks and benchmarks
RISE: Recursive Improvement via Self-Extrapolating Policy DistillationKnowledge distillation → On-policy and self-distillation for LLMs
GUT: Quantifying and Optimizing the Reasoning Uncertainty of LLMs via Graph ComplexityLLM reasoning → Test-time compute and efficient thinking
Testing Interchangeability in LLM Agent TeamsLLM agents and tool use → General LLM agent frameworks and benchmarks
Don't Drop Dropout: Optimizing Layer Sparsity for Efficient LLM Training and InferenceEfficient LLM inference and serving → Early exit and edge inference
How to Speculate about Uncertainty in Agentic Coding? A Draft-Model Gate MethodLLM agents and tool use → Software engineering coding agents
CONTINUITY: Security-Context Contracts for Composable LLM Agent ControlsLLM agents and tool use → Agent threats, guardrails and governance
TacPAC: Tactile Prediction and Real-Time Action Correction in World-Action Models for Contact-Rich ManipulationWorld models → Embodied and robot world models
Trace2Tower: Transition-Aware EigenTrace Induction of Multi-Level Skills for LLM AgentsLLM agents and tool use → Agent skill learning and evolution
Commonsense Reasoning in Computer Vision: Foundations, Recent Advancements, and Future DirectionsLLM reasoning → Multimodal reasoning
Do LLMs Exhibit Coherent Knowledge Structures in Mathematical Reasoning? A Perspective from Knowledge Space TheoryLLM reasoning → Mathematical reasoning
Uncensored Open-weight Models: Redistribution as the Persistence LayerLLM safety and security → Safety alignment, refusal and guardrails
Cross-Domain Tracker Adaptation Without Target-Domain Labels via Vision-Language AgentsMultimodal LLMs / vision-language models → Prompt learning and adaptation of CLIP +1
Substrate-Aware AI Agents: Execution Context as a First-Class InputLLM agents and tool use → General LLM agent frameworks and benchmarks
ACE: Adaptive Calibration-Free Expert Skipping for MoE-based LLMsMixture-of-experts → Expert routing strategies
CABAL: Multi-Agent Simulacra for Tracing the Effects of Collusive Bidding in Peer ReviewLLM agents and tool use → Web, search, and research agents
First Things First: Teaching LLM-Based Agents to Prioritize Must-Haves before Nice-to-HavesLLM agents and tool use → General LLM agent frameworks and benchmarks
A Verifier-Guided Explainable Reasoning Framework with Gold-Anchored QLoRA, Task-Aware Mixture-of-Experts, and Group-Relative RLVRLLM reasoning → Reinforcement learning for reasoning +1
What Matters in On-Policy Distillation? A Perspective on Data Efficiency and Data SelectionLLM reasoning → Prompting and chain-of-thought reasoning +1
LIBERO-RECOVER: Beyond Task Success Towards Failure Recovery in Robotic Manipulation ModelsVision-language-action models → Failure recovery and robustness
WeAgent-MMGenEdit: A Full-Stack Recipe for Multimodal Agentic Image Generation and EditingLLM agents and tool use → Multimodal and embodied agents
Compression Beyond the Uncompressed: A Two-Stage Training Recipe for Soft Context Compression in RAGRetrieval-augmented generation → Generation and context integration +2
From Vision to Language: Investigating Causal Information Flow in Multimodal Decision-MakingMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning
NS-ST-GraphRAG: Neuro-Symbolic Spatio-Temporal GraphRAG for Literary Knowledge ProcessingRetrieval-augmented generation → Knowledge-graph and structured-knowledge RAG
TIER: Threat Implicitness Benchmark for Evaluating LLM Safety BehaviorsLLM safety and security → Safety alignment, refusal and guardrails
Unifying ICL, SFT, KL-Regularized RL Through a Bayesian LensLLM alignment / RLHF / preference optimisation → Reward modeling and RLHF +1
Embedding Surgery: Localized Updates for Adaptive Ranking Correction in Dense RetrievalRetrieval-augmented generation → Generation and context integration
Constructing and Evaluating Clinical Reasoning Trajectories for Medical AgentLLM reasoning → Clinical and biomedical applications
Measuring AI Accountability Through Argumentation Analysis: Can Model Reasoning Withstand Scrutiny?LLM reasoning → Reasoning benchmarks and evaluation
TruthInsightBench: An Evidence-Grounded Benchmark for Automated Evaluation of Open-Ended Scientific Discovery AgentsLLM agents and tool use → Web, search, and research agents
Influence Score and Transformers interpretability: Measure of the Effective Impact of Attention Heads at inference timeLLM safety and security → Prompt injection and agent security
A Structured Debate-Mixture-of-Agents Framework for Complex Clinical Diagnostic Decision SupportLLM agents and tool use → Medical and scientific domain agents
How do LLMs Evaluate Perceived Moral Agency? Investigating Moral Decision-Making in Human-Artificial Agents InteractionsLLM agents and tool use → General LLM agent frameworks and benchmarks
Moral Competence Before Moral Content: Why LLM Agents Lack the Prerequisites for Coherent AlignmentLLM agents and tool use → General LLM agent frameworks and benchmarks
Compositional Reward Models for Conditional Medical Image GenerationLLM alignment / RLHF / preference optimisation → Diffusion and multimodal model alignment
TROVE: Adaptive Agent Skill Orchestration via Trace-Grounded Route Validation and EditingLLM agents and tool use → Agent skill learning and evolution
How a Chatbot's Response Style Shapes a Classroom: A Multi-Agent Simulation of Students Consulting AILLM agents and tool use → Medical and scientific domain agents
Language models judge war differently when tested for alignmentLLM alignment / RLHF / preference optimisation → Human values and AI feedback alignment
BIT.UA at BioASQ 14B: Modular Retrieval with pg_textsearch and Qdrant, and Agent-Based Answer GenerationLLM agents and tool use → Medical and scientific domain agents
A Tree-based RAG Framework for Evidence-Intensive QA via Adaptive Planning and Topology-Aware Evidence GatheringRetrieval-augmented generation → Knowledge-graph and structured-knowledge RAG
BeaconKV: Key-Value Cache Compression Guided by Beacon Queries for Efficient Large Reasoning Model InferenceLLM reasoning → Test-time compute and efficient thinking +1
Fractal basins trap latent reasoningLLM reasoning → Test-time compute and efficient thinking
MCPO: Modality-Contrastive Preference Optimization for Multimodal Chain-of-Thought CompressionLLM reasoning → Multimodal reasoning +2
One Diffusion Model, Two Roles: Guided Trajectory Planning and Safety-Critical Scenario Generation in Closed-Loop SimulationDiffusion models → Diffusion and flow policies for robotics
Compact-Memory LLM Agents via Online Max-Member Clustering and Atom-Aware PackingLLM agents and tool use → Agent memory systems
ARIA - An Agentic Framework for Autonomous Testing of Infotainment SystemsLLM agents and tool use → GUI and mobile agents
TourPhysics: Bringing Physics to World Models for Exploration and Manipulation from a Single ImageWorld models → Latent predictive world models
RefactorPlatform: An Open-Source Harness for Controlled Evaluation of Repository-Scale Refactoring AgentsLLM agents and tool use → Software engineering coding agents
Cache-Aware Joint Router Adaptation for Memory-Efficient MoE InferenceMixture-of-experts → Expert routing strategies
From Language Models to World-Acting Systems: Progress and Limits of Agentic AI across Digital, Social, Virtual, and Physical EnvironmentsLLM agents and tool use → General LLM agent frameworks and benchmarks
Reasoning Without Inference Cost: Latent Semantic Scaffolding for Robot VLA PoliciesLLM reasoning → Spatial and embodied reasoning +1
AtomRec: Evolving Atomic Memory for Agentic RecommendationLLM agents and tool use → Agentic recommender systems
Reinforcement Learning for Sequential Solar PV Policy Design under Uncertainty: An Agent-Based ApproachDeep reinforcement learning methods → RL for energy, networks and operations
Forgetting Without Restarting: Execution-State Unlearning for Stateful LLM AgentsLLM agents and tool use → Agent memory systems +2
From Interaction Traces to Persistent Skills: Online Evolution for Computer-Use AgentsLLM agents and tool use → Agent skill learning and evolution
LLM-Assisted Behavioural and Scenario Augmentation for Agent-Based Energy Adoption ModelsLLM agents and tool use → Financial trading and forecasting agents
CoSkill: Joint Reinforcement Learning of Reasoning and Meta-Skill Agents for Hierarchical Skill EvolutionLLM reasoning → Reasoning for agents, planning and code +1
When Genomic Masking Priors Fail to Transfer: Strong Variant Prediction, Weak Functional GenerationDiffusion models → Flow matching, time series and science

All 401