Papers

New on arXiv

Sat 29 Aug 2026

250 papers

Ingest-Time Fact Compilation for Cost-Efficient and Reliable Question Answering over Revised CorporaLLM agents and tool use → Agent memory systems
AgentKernel: The Trust-Native Agentic Operating SystemLLM agents and tool use → Agent threats, guardrails and governance
Free the Language Model From the Vision Encoder: Semantic Serialization as a Perception Interface for Small Language ModelsLLM reasoning → Multimodal reasoning +1
A Pinch of SFT, A Dash of RL: When Reinforcement Learning Helps Long-Horizon Advertising AgentsDeep reinforcement learning methods → RL training of language and agent models
Evidence-Order Calibration for Selective Visual Reasoning under Progressive Loss of Question-Critical EvidenceLLM reasoning → Multimodal reasoning +1
trajectory-judge: What Outcome-Only LLM Judges Miss on Agent TrajectoriesLLM agents and tool use → General LLM agent frameworks and benchmarks
SilentProbe: Measuring Silent Failure in Production APIs Used as Agent ToolsLLM agents and tool use → Tool learning and function calling
Knowledge Distillation under Teacher Misspecification: An Order-Parameter Analysis of the Gap between Teacher Mimicry and Task PerformanceKnowledge distillation → Knowledge distillation for LLMs
Chain-of-Thought Faithfulness of Reasoning Models Varies with Where and How Preference Cues Are DeliveredLLM reasoning → Prompting and chain-of-thought reasoning
Can escalation channels redirect reward hacking toward defect disclosure?LLM agents and tool use → Software engineering coding agents +1
Text-Guided Diffusion-Based Adversarial Attacks on Chest X-Ray ImagesDiffusion models → Medical imaging and segmentation
Does Latent Planning Survive Point Clouds? Action-Conditioned JEPA World Models for Geometric ObservationsWorld models → Latent predictive world models
SMILE: Smooth Motion for Improved Long-Horizon VLA ExecutionVision-language-action models → Efficient VLA inference and adaptation
Evaluating the Semantic Specificity of Representation Steering in Language ModelsLLM reasoning → Reasoning benchmarks and evaluation +1
Where Induction Runs Out: Description-Length Difficulty and the Memorisation Gap in Integer-Sequence BenchmarksLLM reasoning → Mathematical reasoning
Agents as Knowledge Integrator and Utilizer in Multimodal RecommendationLLM agents and tool use → Agentic recommender systems
GATE: Reliability-Gated Gaussian Evidence Fusion for Training-Free Test-Time Adaptation of Vision-Language ModelsMultimodal LLMs / vision-language models → Prompt learning and adaptation of CLIP
Arabic Safety Alignment as Selective Refusal: An Empirical Study of SFT, DPO, and Guard CalibrationLLM alignment / RLHF / preference optimisation → Safety alignment and jailbreak attacks +1
Think, Look, and Revise: Inconsistency-Aware Visual Self-Correction in MLLMsLLM reasoning → Multimodal reasoning +1
Unlearning on Spatio-Temporal Graphs through Subgraph Virtual Edge ReconstructionLLM safety and security → Graph unlearning
TRACER: Per-Tool Context Retention for LLM Agents via Consequence-Attributed Reinforcement LearningLLM reasoning → Reasoning for agents, planning and code +1
Spatial Entropy based Partitioning for Spatiotemporal Graph UnlearningLLM safety and security → Graph unlearning
LiteSearch-VL: Small Multimodal Search Agents via Trajectory Distillation and Synthetic Step-DPOLLM agents and tool use → Multimodal and embodied agents +1
Cross-Relational Preference Learning for Better LLM Instruction FollowingLLM alignment / RLHF / preference optimisation → LLM preference optimization methods
StageWell: A Process-Aligned Chinese Corpus for Positive-Psychology Support DialogueLLM alignment / RLHF / preference optimisation → Medical LLM alignment
Test-Time Scaling for Video Diffusion Models via Diagnosis-Guided Candidate RecyclingLLM reasoning → Diffusion reasoning and test-time scaling +1
Hyper3-CLIP: Hierarchy-Conditioned Hyperbolic Vision-Language TrainingMultimodal LLMs / vision-language models → Prompt learning and adaptation of CLIP
Super Library Agent: Joint Generation and Maintenance of Multiple Applications Beyond the Single CodebaseLLM agents and tool use → Software engineering coding agents
Evaluating LLM-based AI agents integrated with materials synthesis tools: the case of atomic layer depositionLLM agents and tool use → Medical and scientific domain agents
Detecting and Repairing Hallucinations in Retrieval-Augmented GenerationRetrieval-augmented generation → Hallucination detection in RAG
Learning Simple Test-Time Environments for LLM Web AgentsLLM agents and tool use → Web, search, and research agents
Accelerating Unified Multimodal Models with Core-Expansion Routing and Unified Computation SchedulingEfficient LLM inference and serving → Visual token pruning in multimodal models
MMPCBench: Benchmarking Multimodal Large Language Models on Proactive Critique of Flawed InputsMultimodal LLMs / vision-language models → MLLM training and visual instruction tuning
3D-MRL: Nested Multimodal 3D Representations via Matryoshka Representation LearningMultimodal LLMs / vision-language models → Embodied and 3D vision-language models
Elastic Token Compression for Pixel-Space Diffusion TransformersDiffusion models → Diffusion foundations and sampling +1
Modality Fault Lines: Structural Corruptions Reveal Fragile Omni-Modal ReasoningLLM reasoning → Multimodal reasoning
Learning to Ground Before Reading: Unified PCB Engineering Drawing Parsing with Compact Vision-Language ModelsMultimodal LLMs / vision-language models → Chart understanding and anomaly detection
RACER: Reinforced Agent Collaboration for Explainable Reasoning on Knowledge GraphsLLM reasoning → Knowledge- and tool-augmented reasoning +1
Sense Once, Serve Many: Common-Trace Factorized Constrained PPO for Online Sensing-Session Consolidation in Multi-Tenant ISAC NetworksDeep reinforcement learning methods → RL for energy, networks and operations
Dynamic Important Example Mining for Reinforcement FinetuningLLM reasoning → Reinforcement learning for reasoning
RL-FAT: Reinforcement Learning for Fair Adversarial TrainingDeep reinforcement learning methods → Adversarial robustness and RL in finance
DARD: Zero-Shot Degradation-Aware Retinex-Guided Diffusion for Low-Light Image EnhancementDiffusion models → Inverse problems and image restoration
AnyWorld: Factorized Egocentric World Models for Cross-Embodiment GeneralizationWorld models → Embodied and robot world models
Localizing Emergent Failures in Agentic AI: Recovering Minimal Repair Families via Counterfactual ReplayLLM agents and tool use → Tool learning and function calling
Attribute-Based Activation Steering of LLMs for Group-Specific Explanation GenerationMechanistic interpretability → Activation steering methods
AdaVLA: Adaptive Step Flow Matching for Training-free Acceleration of Vision-Language-Action ModelsLLM reasoning → Spatial and embodied reasoning +3
Benevolent Bias in Multi-Turn Human-Agent DialogueLLM agents and tool use → Social simulation and multi-agent debate
AgentLogs: A Dataset for Opening the Black Box of GitHub's Cloud AgentLLM agents and tool use → Software engineering coding agents
HalluPrism: When Multimodal Uncertainty Should Diagnose, Not DecideMultimodal LLMs / vision-language models → MLLM training and visual instruction tuning
Locked at the Entrance, Open Inside: Where RLVR Narrows the Solution SpaceLLM reasoning → Reinforcement learning for reasoning
Multi-Scale Temporal Domain Alignment for Federated Video Domain AdaptationKnowledge distillation → General knowledge distillation methods
Foundational feature fusion for conditional flow matching in 6D pose estimationDiffusion models → 3D, point cloud and scene generation
JudgePanel: A Compact Judge with Panel Deliberation via Adaptive Multi-Reward Reinforcement LearningDeep reinforcement learning methods → RL training of language and agent models
Training-Free Hidden-State Refinement for Flow-Matching Image GeneratorsDiffusion models → Flow matching, time series and science
Manifold-Constrained PET Reconstruction with Learned Flow-Matching PriorsDiffusion models → Medical imaging and segmentation
WoE Wrote It? Watermarking Mixture-of-Experts LLMs for Black-Box Text ProvenanceLLM safety and security → Watermarking generated content and models
Quantifying Error Tolerance in Synthetic Data: An Atomic-level Operand vs. Operator Perturbation StudySynthetic data and self-improvement for LLMs → LLM synthetic data generation
Chat-Edit-3D++: Interactive 3D and 4D Scene Editing via Large Language ModelsMultimodal LLMs / vision-language models → Embodied and 3D vision-language models
APIFlow-Bench: Measuring Whether Agents Survive Long, Dependent API WorkflowsLLM agents and tool use → Software engineering coding agents
HEAR Who Said What: Unlocking Speaker-Attributed Reasoning via Counterfactual Voice GroundingLLM reasoning → Audio-language model reasoning
Temperature-Adaptive Transformed Teacher MatchingKnowledge distillation → General knowledge distillation methods
Development of an Autonomous AI Coding Agent using Monte Carlo Tree Search (MCTS) and Gemini LLM FrameworksLLM agents and tool use → Software engineering coding agents
Titans-QFWP: A Regime-Aware Hybrid Quantum Fast Weight Programmer for Portfolio OptimizationDeep reinforcement learning methods → Adversarial robustness and RL in finance
EviAnchor: Mitigating Hallucinations in Large Vision-Language Models via Regional Visual Evidence CompensationMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning
DREAM: Deployment-Time Demonstration Generation via Real-to-Sim for Scalable Policy AdaptationVision-language-action models → Reinforcement learning and post-training
Revolutionizing Turn-by-Turn Navigation with Cloud-Edge Deep LearningMixture-of-experts → MoE in reinforcement learning and robotics
Selective Disclosure of Hidden Directives in Reasoning Models: Behavioral Asymmetry and SteeringLLM reasoning → Prompting and chain-of-thought reasoning
Agent2UCB: Agentic System for Generative Engine OptimizationLLM agents and tool use → Web, search, and research agents
PathBridger: Subgoal Bridges for Offline Goal-Conditioned Reinforcement LearningOffline RL and imitation learning from data → Diffusion and flow policies for offline RL
DReSG: Diffusion Residuals for Stylized Gaussian SplattingDiffusion models → 3D, point cloud and scene generation
DocIntent: Answerability-Guided Agentic Restoration for Real-World Document Visual Question AnsweringMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning +1
A Unifying Perspective on Language Model Representations: From Filler-Role Structure to Mechanistic InterpretabilityMechanistic interpretability → Knowledge and representations inside LLMs
Learning to Follow In-Context Watermark Instructions via Self-DistillationLLM safety and security → Watermarking generated content and models +1
Flow-JEPA: Flow Matching for Robust Latent Dynamics in JEPA World ModelsWorld models → Latent predictive world models +1
Facts Without Rules: Boundary Metadata Collapse in Multi-Agent LLM HandoffsLLM agents and tool use → Agent threats, guardrails and governance
A Degradation-Tolerance Benchmark for Camera-Only End-to-End DrivingEnd-to-end autonomous driving → End-to-end system architecture
Hydra: A Navigation World Action Model with Discrete Latent Planning and Continuous Flow-Matching ExecutionWorld models → Embodied and robot world models +1
Selective Forgetting: A Graph-Based Memory Framework for Long-Term LLM AgentsRetrieval-augmented generation → Agentic RAG, agents and memory +1
Legacy System Modernization with Coding Agents: A Case StudyLLM agents and tool use → Software engineering coding agents
Continuity-Free Near-Minimax Leading-Order Regret for CVaR-UCBVIDeep reinforcement learning methods → RL theory and sample complexity
CoVA-SFT: A Large-Scale Dataset for Chain of Visual AbstractionsLLM reasoning → Multimodal reasoning
How To Do Things With Prompts
VG-TIE: An interpretable tabular-to-image encoding method based on visibility graphs
Albireo: Adaptive, Energy-Efficient Inference Framework for Video Object Detection on the Edge
SpectralCTGaussians: Projection-Domain Reconstruction and Basis Material Decomposition for Spectral CT using 3D Gaussian Splatting
A Manifold-Aware Topic Modeling Approach via Rank-Based Prototypes
Optimal Recovery Meets Bayesian Learning: Where Worst-Case Bounds Pay Off
SegTSim: A Big Data Driven Segmented Temporal Simulation Framework for Heterogeneous Multivariate Systems
WildfireSpreadBench: The Metric Decides the Model in Wildfire Spread Prediction
Cortex: Content Analysis Support Software, a Resource for Qualitative Research
Reliable iToF Depth Sensing via Sensor-Intrinsic Uncertainty Modeling and State-Space Restoration
ECHO: Dyadic 3D Facial Motion Generation with Asymmetric Deterministic Articulation and Stochastic Reaction
SciLitBench: Benchmark and Design Principles for LLM-Powered Systematic Literature Reviews
Efficient Context-Limited Telescope Bibliography Classification for the WASP-2025 Shared Task Using SciBERT
A Globally Convergent Algorithm for Total Scaled-Gradient Variation via Cone-Constrained Bilinear Decomposition
Co-Evolutionary Prompt Optimization with Cross-Category Transfer for Zero-Shot Anomaly Detection
Deciding When to Decide: Testing Operational Suboptimality Under Distributional Shift
Benchmark Contamination: A Taxonomy Organized by Defeated Mitigation
A Causal Model for Locating and Unlocking Sandbagging in Model Organisms
Toward Latent Language Model Skills Steering and Optimization: An Empirical Study

All 250