Papers

New on arXiv

Wed 11 Feb 2026

460 papers

MATRAG: Multi-Agent Transparent Retrieval-Augmented Generation for Explainable RecommendationsRetrieval-augmented generation → RAG for recommendation +1
StreamServe: Adaptive Speculative Flows for Low-Latency Disaggregated LLM ServingEfficient LLM inference and serving → Serving systems and request scheduling
DRAFT: Task Decoupled Latent Reasoning for Agent SafetyLLM reasoning → Reasoning for agents, planning and code +1
UI-Oceanus: Scaling GUI Agents with Synthetic Environmental DynamicsLLM agents and tool use → GUI and mobile agents +1
Prompt-Dependent Ranking of Large Language Models with Uncertainty QuantificationLLM alignment / RLHF / preference optimisation → LLM preference optimization methods
Compressed Sensing for Capability Localization in Large Language ModelsLLM reasoning → Test-time compute and efficient thinking
The CompMath-MCQ Dataset: Are LLMs Ready for Higher-Level Math?LLM reasoning → Mathematical reasoning
Training-free Dropout Sampling for Semantic Token Acceptance in Speculative DecodingEfficient LLM inference and serving → Speculative decoding
Fragile Thoughts: How Large Language Models Handle Chain-of-Thought PerturbationsLLM reasoning → Prompting and chain-of-thought reasoning
Neural Paging: Learning Context Management Policies for Turing-Complete AgentsLLM agents and tool use → Agent memory systems
Scaling Reward Modeling without Human SupervisionLLM alignment / RLHF / preference optimisation → Reward modeling and RLHF
APEX-SQL: Talking to the data via Agentic Exploration for Text-to-SQLLLM agents and tool use → Web, search, and research agents
Understand Then Memory: A Cognitive Gist-Driven RAG Framework with Global Semantic DiffusionRetrieval-augmented generation → Knowledge-graph and structured-knowledge RAG
Beyond Context Sharing: A Unified Agent Communication Protocol (ACP) for Secure, Federated, and Autonomous Agent-to-Agent (A2A) OrchestrationLLM agents and tool use → Agent threats, guardrails and governance
MedScope: Incentivizing "Think with Videos" for Clinical Reasoning via Coarse-to-Fine Tool CallingLLM reasoning → Clinical and biomedical applications
HiST-VLA: A Hierarchical Spatio-Temporal Vision-Language-Action Model for End-to-End Autonomous DrivingVision-language-action models → VLA for autonomous driving +1
Grandes Modelos de Linguagem Multimodais (MLLMs): Da Teoria `a Pr'aticaMultimodal LLMs / vision-language models → MLLM training and visual instruction tuning
LoopFormer: Elastic-Depth Looped Transformers for Latent Reasoning via Shortcut ModulationLLM reasoning → Test-time compute and efficient thinking
Filtered Approximate Nearest Neighbor Search in Vector Databases: System Design and Performance AnalysisRetrieval-augmented generation → Vector search and ANN indexes
Optimizing Agent Planning for Security and AutonomyLLM safety and security → Prompt injection and agent security
When Visibility Outpaces Verification: Delayed Verification and Narrative Lock-in in Agentic AI DiscourseLLM agents and tool use → Social simulation and multi-agent debate
TRACER: Trajectory Risk Aggregation for Critical Episodes in Agentic ReasoningLLM agents and tool use → General LLM agent frameworks and benchmarks
Latent Forcing: Reordering the Diffusion Trajectory for Pixel-Space Image GenerationDiffusion models → Diffusion foundations and sampling
Can We Really Learn One Representation to Optimize All Rewards?Deep reinforcement learning methods → Deep RL algorithms and representations
General and Efficient Steering of Diffusion ModelsDiffusion models → Diffusion foundations and sampling
Causal-JEPA: Learning World Models through Object-Level Latent MaskingWorld models → Latent predictive world models
Sparse Semantic Dimension as a Generalization Certificate for LLMsMechanistic interpretability → SAE features and their applications
The Energy of Falsehood: Detecting Hallucinations via Diffusion Model LikelihoodsDiffusion models → Medical imaging and segmentation
Finding the Cracks: Improving LLMs Reasoning with Paraphrastic Probing and Consistency VerificationLLM reasoning → Test-time compute and efficient thinking
ReplicatorBench: Benchmarking LLM Agents for Replicability in Social and Behavioral SciencesLLM agents and tool use → Web, search, and research agents
Pushing Forward Pareto Frontiers of Proactive Agents with Behavioral Agentic OptimizationLLM agents and tool use → Reinforcement learning for LLM agents
AgentNoiseBench: Benchmarking Robustness of Tool-Using LLM Agents Under Noisy ConditionLLM agents and tool use → General LLM agent frameworks and benchmarks
Bi-Level Prompt Optimization for Multimodal LLM-as-a-JudgeMultimodal LLMs / vision-language models → MLLM training and visual instruction tuning
Evaluating Alignment of Behavioral Dispositions in LLMsLLM alignment / RLHF / preference optimisation → Human values and AI feedback alignment
Security Threat Modeling for Emerging AI-Agent Protocols: A Comparative Analysis of MCP, A2A, Agora, and ANPLLM agents and tool use → Agent threats, guardrails and governance
Are Aligned Large Language Models Still Misaligned?LLM alignment / RLHF / preference optimisation → LLM preference optimization methods
CryptoAnalystBench: Failures in Multi-Tool Long-Form LLM AnalysisLLM agents and tool use → Agent threats, guardrails and governance
The PBSAI Governance Ecosystem: A Multi-Agent AI Reference Architecture for Securing Enterprise AI EstatesLLM agents and tool use → Agent threats, guardrails and governance
H-WM: Robotic Task and Motion Planning Guided by Hierarchical World ModelWorld models → Embodied and robot world models
Peak + Accumulation: A Proxy-Level Scoring Formula for Multi-Turn LLM Attack DetectionLLM safety and security → Jailbreaking
Evaluating Memory Structure in LLM AgentsRetrieval-augmented generation → Agentic RAG, agents and memory +1
Active Zero: Self-Evolving Vision-Language Models through Active Environment ExplorationMultimodal LLMs / vision-language models → Embodied and 3D vision-language models +1
SurveyLens: A Discipline-Aware Benchmark for Automatic Survey GenerationLLM agents and tool use → Web, search, and research agents
Agent-Diff: Benchmarking LLM Agents on Enterprise API Tasks via Code Execution with State-Diff-Based EvaluationLLM agents and tool use → Software engineering coding agents
Patch the Distribution Mismatch: RL Rewriting Agent for Stable Off-Policy SFTLLM agents and tool use → Web, search, and research agents
Charting Empirical Laws for LLM Fine-Tuning in Scientific Multi-Discipline LearningMixture-of-experts → MoE large language models
Transferable Backdoor Attacks for Code Models via Sharpness-Aware Adversarial PerturbationLLM safety and security → Backdoor and data poisoning attacks
SWE-MiniSandbox: Container-Free Reinforcement Learning for Building Software Engineering AgentsLLM agents and tool use → Software engineering coding agents +1
Data Repetition Beats Data Scaling in Long-CoT Supervised Fine-TuningLLM reasoning → Test-time compute and efficient thinking
Beyond VLM-Based Rewards: Diffusion-Native Latent Reward ModelingDiffusion models → Diffusion and flow policies for robotics +1
APEX: Learning Adaptive High-Platform Traversal for Humanoid RobotsDeep reinforcement learning methods → RL for robot locomotion and manipulation
Data-Efficient Hierarchical Goal-Conditioned Reinforcement Learning via Normalizing FlowsDeep reinforcement learning methods → Deep RL algorithms and representations
FormalJudge: A Neuro-Symbolic Paradigm for Agentic OversightLLM agents and tool use → General LLM agent frameworks and benchmarks
Just on Time: Token-Level Early Stopping for Diffusion Language ModelsDiffusion models → Discrete and text diffusion
Meltdown: Circuits and Bifurcations in Point-Cloud-Conditioned 3D Diffusion TransformersDiffusion models → 3D, point cloud and scene generation
Asymmetric Prompt Weighting for Reinforcement Learning with Verifiable RewardsLLM reasoning → Reinforcement learning for reasoning
PhyCritic: Multimodal Critic Models for Physical AIMultimodal LLMs / vision-language models → Embodied and 3D vision-language models
HairWeaver: Few-Shot Photorealistic Hair Motion Synthesis with Sim-to-Real Guided Video DiffusionDiffusion models → 3D, point cloud and scene generation
Learning to Compose for Cross-domain Agentic Workflow GenerationLLM agents and tool use → Software engineering coding agents
FastFlow: Accelerating The Generative Flow Matching Models with Bandit InferenceDiffusion models → Flow matching, time series and science
GameDevBench: Evaluating Agentic Capabilities Through Game DevelopmentLLM agents and tool use → Software engineering coding agents
Safety Recovery in Reasoning Models Is Only a Few Early Steering Steps AwayLLM reasoning → Reasoning for agents, planning and code +1
General Flexible f-divergence for Challenging Offline RL Datasets with Low Stochasticity and Diverse Behavior PoliciesOffline RL and imitation learning from data → Offline RL algorithms
SteuerLLM: Local specialized large language model for German tax law analysisLLM reasoning → Reasoning for agents, planning and code
RISE: Self-Improving Robot Policy with Compositional World ModelWorld models → Embodied and robot world models +2
Chatting with Images for Introspective Visual ThinkingLLM reasoning → Multimodal reasoning +1
S-MARC: Causal Streaming Reasoning for Full-Duplex Conversational Behavior ModelingLLM reasoning → Audio-language model reasoning
A Gibbs posterior sampler for inverse problem based on prior diffusion modelDiffusion models → Inverse problems and image restoration
Embedding Inversion via Conditional Masked Diffusion Language ModelsDiffusion models → Discrete and text diffusion
Reality Copilot: Voice-First Human-AI Collaboration in Mixed Reality Using Large Multimodal ModelsMultimodal LLMs / vision-language models → MLLM training and visual instruction tuning
Chain-of-Look Spatial Reasoning for Dense Surgical Instrument CountingLLM reasoning → Clinical and biomedical applications
ContactGaussian-WM: Learning Physics-Grounded World Model from VideosWorld models → Latent predictive world models
OSIL: Learning Offline Safe Imitation Policies with Safety Inferred from Non-preferred TrajectoriesOffline RL and imitation learning from data → Safe offline RL
CLI-Gym: Scalable CLI Task Generation via Agentic Environment InversionLLM agents and tool use → Software engineering coding agents
Multi-Task Reinforcement Learning of Drone Aerobatics by Exploiting Geometric SymmetriesDeep reinforcement learning methods → RL for robot navigation and UAV control
TVCACHE: A Stateful Tool-Value Cache for Post-Training LLM AgentsLLM agents and tool use → Software engineering coding agents
Scaling World Model for Hierarchical Manipulation PoliciesWorld models → Embodied and robot world models +1
RADAR: Benchmarking Vision-Language-Action Generalization via Real-World Dynamics, Spatial-Physical Intelligence, and Autonomous EvaluationVision-language-action models → Generalist VLA models and benchmarks
FeatureBench: Benchmarking Agentic Coding for Complex Feature DevelopmentLLM agents and tool use → Software engineering coding agents
MoEEdit: Efficient and Routing-Stable Knowledge Editing for Mixture-of-Experts LLMsMixture-of-experts → MoE large language models
Search or Accelerate: Confidence-Switched Position Beam Search for Diffusion Language ModelsDiffusion models → Discrete and text diffusion
FastUSP: A Multi-Level Collaborative Acceleration Framework for Distributed Diffusion Model InferenceDiffusion models → Diffusion foundations and sampling
Near-Constant Strong Violation and Last-Iterate Convergence for Online CMDPs via Decaying Safety MarginsDeep reinforcement learning methods → RL theory and sample complexity
Blind Gods and Broken Screens: Architecting a Secure, Intent-Centric Mobile Agent Operating SystemLLM agents and tool use → Agent threats, guardrails and governance
Revisiting Regularized Policy Optimization for Stable and Efficient Reinforcement Learning in Two-Player GamesDeep reinforcement learning methods → RL for games and combinatorial optimization
Reinforcing Chain-of-Thought Reasoning with Self-Evolving RubricsLLM reasoning → Prompting and chain-of-thought reasoning +2
ResWorld: Temporal Residual World Model for End-to-End Autonomous DrivingWorld models → World models for autonomous driving +1
Chart Specification: Structural Representations for Incentivizing VLM Reasoning in Chart-to-Code GenerationLLM reasoning → Multimodal reasoning +1
Agentic Knowledge Distillation: Autonomous Training of Small Language Models for SMS Threat DetectionLLM agents and tool use → Agent threats, guardrails and governance
ICA: Information-Aware Credit Assignment for Visually Grounded Long-Horizon Information-Seeking AgentsDeep reinforcement learning methods → RL training of language and agent models
Training and Benchmarking Code Generation for Physics-Inspired AnimationsLLM reasoning → Spatial and embodied reasoning
Mirror descent actor-critic methods for entropy regularised MDPs in general spaces: stability and convergenceDeep reinforcement learning methods → RL theory and sample complexity
RePO: Bridging On-Policy Learning and Off-Policy Knowledge through Rephrasing Policy OptimizationLLM alignment / RLHF / preference optimisation → Reward modeling and RLHF
Beyond Confidence: The Rhythms of Reasoning in Generative ModelsLLM reasoning → Test-time compute and efficient thinking
Why Does RL Generalize Better Than SFT? A Data-Centric Perspective on VLM Post-TrainingMultimodal LLMs / vision-language models → Embodied and 3D vision-language models
See, Plan, Snap: Evaluating Multimodal GUI Agents in ScratchLLM agents and tool use → GUI and mobile agents
DeepImageSearch: Benchmarking Multimodal Agents for Context-Aware Image Retrieval in Visual HistoriesLLM agents and tool use → Multimodal and embodied agents
RSHallu: Dual-Mode Hallucination Evaluation for Remote-Sensing Multimodal Large Language Models with Domain-Tailored MitigationMultimodal LLMs / vision-language models → Hallucination mitigation in LVLMs
Transport, Don't Generate: Deterministic Geometric Flows for Combinatorial OptimizationDiffusion models → Diffusion foundations and sampling
Semi-Supervised Cross-Domain Imitation LearningOffline RL and imitation learning from data → Imitation learning methods

All 460