Papers

New on arXiv

Tue 7 Apr 2026

498 papers

What is Good? Extracting and Testing Implicit Theories of Literary Quality from LLM Reasoning TracesLLM reasoning → Reasoning benchmarks and evaluation
MechELK: A Mechanistic Interpretability Framework for Eliciting Latent Knowledge in Large Language ModelsLLM reasoning → Reasoning benchmarks and evaluation +1
What are They Thinking? Delineation, Probing, and Tracking of Concepts in LLMsLLM reasoning → Reasoning benchmarks and evaluation
Cross-Platform Fused MoE Dispatch in Triton: Portable Expert Routing Without CUDAMixture-of-experts → Efficient MoE training and inference systems
RealUserSim: Bridging the Reality Gap in Agent Benchmarking via Grounded User SimulationLLM agents and tool use → Social simulation and multi-agent debate
Stress-Testing the Reasoning Competence of LLMs With Proofs Under Minimal FormalismLLM reasoning → Mathematical reasoning
Reasoning-Guided Grounding: Elevating Video Anomaly Detection through Multimodal Large Language ModelsLLM reasoning → Chain-of-thought reasoning for driving +1
Agentic AI-Based Joint Computing and Networking via Mixture of Experts and Large Language ModelsLLM agents and tool use → Agentic AI for wireless networks +1
Understanding Emergent Misalignment via Feature Superposition GeometryMechanistic interpretability → Steering and probing for LLM safety
LLM BiasesLLM agents and tool use → Agentic recommender systems
CareGuardAI: Context-Aware Multi-Agent Guardrails for Clinical Safety & Hallucination Mitigation in Patient-Facing LLMsLLM agents and tool use → Medical and scientific domain agents +1
SkillGraph: Graph Foundation Priors for LLM Agent Tool Sequence RecommendationLLM agents and tool use → Tool learning and function calling
SynopticBench: Evaluating Vision-Language Models on Generating Weather Forecast Discussions of the FutureMultimodal LLMs / vision-language models → Embodied and 3D vision-language models
Multi-Granularity Reasoning for Image Quality Assessment via Attribute-Aware Reinforcement Learning to RankLLM reasoning → Multimodal reasoning +1
Attention-Guided Flow-Matching for Sparse 3D Geological GenerationDiffusion models → Flow matching, time series and science
OpenKedge: Governing Agentic Mutation with Execution-Bound Safety and Evidence ChainsLLM agents and tool use → Agent threats, guardrails and governance
Constrained Policy Optimization for Provably Fair Order MatchingDeep reinforcement learning methods → Adversarial robustness and RL in finance
Efficient Quantization of Mixture-of-Experts with Theoretical Generalization GuaranteesMixture-of-experts → MoE large language models
MedConclusion: A Benchmark for Biomedical Conclusion Generation from Structured AbstractsLLM reasoning → Clinical and biomedical applications
VLMShield: Efficient and Robust Defense of Vision-Language Models against Malicious PromptsMultimodal LLMs / vision-language models → MLLM jailbreaks and adversarial robustness
Transformer See, Transformer Do: Copying as an Intermediate Step in Learning Analogical ReasoningLLM reasoning → Reasoning benchmarks and evaluation
Hyperfastrl: Hypernetwork-based reinforcement learning for unified control of parametric chaotic PDEsDeep reinforcement learning methods → RL theory and sample complexity
Improving Robustness In Sparse Autoencoders via Masked RegularizationMechanistic interpretability → Sparse autoencoder training and evaluation
Discrete Flow Matching Policy OptimizationDiffusion models → Diffusion and flow policies for robotics +1
Distributed Interpretability and Control for Large Language ModelsMechanistic interpretability → Activation steering methods
Multi-objective Evolutionary Merging Enables Efficient Reasoning ModelsLLM reasoning → Test-time compute and efficient thinking
A Control Barrier Function-Constrained Model Predictive Control Framework for Safe Reinforcement LearningDeep reinforcement learning methods → Safe and constrained RL
Learning to Interrupt in Language-based Multi-agent CommunicationLLM agents and tool use → General LLM agent frameworks and benchmarks
The Defense Trilemma: Why Prompt Injection Defense Wrappers Fail?LLM safety and security → Prompt injection and agent security
The Depth Ceiling: On the Limits of Large Language Models in Discovering Latent PlanningLLM reasoning → Prompting and chain-of-thought reasoning
When to Call an Apple Red: Humans Follow Introspective Rules, VLMs Don'tMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning
State-of-the-Art Arabic Language Modeling with Sparse MoE Fine-Tuning and Chain-of-Thought DistillationLLM reasoning → Reasoning benchmarks and evaluation +1
ODE-free Neural Flow Matching for One-Step Generative ModelingDiffusion models → Flow matching, time series and science
ProofSketcher: Hybrid LLM + Lightweight Proof Checker for Reliable Math/Logic ReasoningLLM reasoning → Mathematical reasoning
Qualixar OS: A Universal Operating System for AI Agent OrchestrationLLM agents and tool use → General LLM agent frameworks and benchmarks
SELFDOUBT: Uncertainty Quantification for Reasoning LLMs via the Hedge-to-Verify RatioLLM reasoning → Test-time compute and efficient thinking
The Master Key Hypothesis: Unlocking Cross-Model Capability Transfer via Linear Subspace AlignmentLLM reasoning → Mathematical reasoning
MTA-Agent: An Open Recipe for Multimodal Deep Search AgentsMultimodal LLMs / vision-language models → Document understanding and multimodal RAG +1
SymptomWise: A Deterministic Reasoning Layer for Reliable and Efficient AI SystemsLLM reasoning → Clinical and biomedical applications
The Illusion of Superposition? A Principled Analysis of Latent Thinking in Language ModelsLLM reasoning → Prompting and chain-of-thought reasoning
Beyond Functional Correctness: Design Issues in AI IDE-Generated Large-Scale ProjectsLLM agents and tool use → Software engineering coding agents
ForkKV: Scaling Multi-LoRA Agent Serving via Copy-on-Write Disaggregated KV CacheLLM agents and tool use → Software engineering coding agents +1
WebSP-Eval: Evaluating Web Agents on Website Security and Privacy TasksLLM agents and tool use → Agent threats, guardrails and governance
Evidence-Based Actor-Verifier Reasoning for Echocardiographic AgentsLLM reasoning → Clinical and biomedical applications
Evolution of Video Generative FoundationsWorld models → Video-generation world models and simulators +1
A Novel Automatic Framework for Speaker Drift Detection in Synthesized SpeechDiffusion models → Audio, speech and music
The End of Human Judgment in the Kill Chain? Relocating Initiative and Interpretation with Agentic AILLM agents and tool use → General LLM agent frameworks and benchmarks
Limits of Difficulty Scaling: Hard Samples Yield Diminishing Returns in GRPO-Tuned SLMsLLM reasoning → Reinforcement learning for reasoning +1
AgentOpt v0.1 Technical Report: Client-Side Optimization for LLM-Based AgentLLM agents and tool use → Software engineering coding agents
TalkLoRA: Communication-Aware Mixture of Low-Rank Adaptation for Large Language ModelsMixture-of-experts → Mixture of LoRA experts
Harnessing Hyperbolic Geometry for Harmful Prompt Detection and SanitizationMultimodal LLMs / vision-language models → MLLM jailbreaks and adversarial robustness +1
ClawLess: A Security Model of AI AgentsLLM agents and tool use → Agent threats, guardrails and governance
MAT-Cell: A Multi-Agent Tree-Structured Reasoning Framework for Batch-Level Single-Cell AnnotationLLM reasoning → Clinical and biomedical applications +1
RAGEN-2: Reasoning Collapse in Agentic RLLLM reasoning → Reinforcement learning for reasoning +1
ToxReason: A Benchmark for Mechanistic Chemical Toxicity Reasoning via Adverse Outcome PathwayLLM reasoning → Clinical and biomedical applications
S3: Stratified Scaling Search for Test-Time in Diffusion Language ModelsLLM reasoning → Diffusion reasoning and test-time scaling +1
Paper Circle: An Open-source Multi-agent Research Discovery and Analysis FrameworkLLM agents and tool use → Web, search, and research agents
Action Images: End-to-End Policy Learning via Multiview Video GenerationWorld models → Embodied and robot world models
HaloProbe: Bayesian Detection and Mitigation of Object Hallucinations in Vision-Language ModelsMultimodal LLMs / vision-language models → Hallucination mitigation in LVLMs
DiffHDR: Re-Exposing LDR Videos with Video Diffusion ModelsDiffusion models → Inverse problems and image restoration
MMEmb-R1: Reasoning-Enhanced Multimodal Embedding with Pair-Aware Selection and Adaptive ControlLLM reasoning → Multimodal reasoning
Toward Consistent World Models with Multi-Token Prediction and Latent Semantic EnhancementWorld models → World models for language agents
Exclusive UnlearningLLM safety and security → Machine unlearning
Claw-Eval: Towards Trustworthy Evaluation of Autonomous AgentsLLM agents and tool use → General LLM agent frameworks and benchmarks
Gym-Anything: Turn any Software into an Agent EnvironmentLLM agents and tool use → GUI and mobile agents
Lightweight Multimodal Adaptation of Vision Language Models for Species Recognition and Habitat Context Interpretation in Drone Thermal ImageryMultimodal LLMs / vision-language models → Embodied and 3D vision-language models
SEM-ROVER: Semantic Voxel-Guided Diffusion for Large-Scale Driving Scene GenerationDiffusion models → Diffusion and flow policies for robotics
Masking or Mitigating? Deconstructing the Impact of Query Rewriting on Retriever Biases in RAGRetrieval-augmented generation → RAG robustness and adaptive retrieval
Scientific Graphics Program Synthesis via Dual Self-Consistency Reinforcement LearningLLM reasoning → Multimodal reasoning +1
Short Data, Long Context: Distilling Positional Knowledge in TransformersKnowledge distillation → Knowledge distillation for LLMs
HiPolicy: Hierarchical Multi-Frequency Action Chunking for Policy LearningOffline RL and imitation learning from data → Robot manipulation imitation learning
From Hallucination to Structure Snowballing: The Alignment Tax of Constrained Decoding in LLM ReflectionLLM reasoning → Test-time compute and efficient thinking +1
Lipschitz regularity in Flow Matching and Diffusion Models: sharp sampling rates and functional inequalitiesDiffusion models → Flow matching, time series and science
Value Mirror Descent for Reinforcement LearningDeep reinforcement learning methods → RL theory and sample complexity
CodecSight: Leveraging Video Codec Signals for Efficient Streaming VLM InferenceMultimodal LLMs / vision-language models → Visual token pruning and efficiency
BiMind: A Dual-Head Reasoning Model with Attention-Geometry Adapter for Incorrect Information DetectionLLM reasoning → Test-time compute and efficient thinking
Epistemic Blinding: An Inference-Time Protocol for Auditing Prior Contamination in LLM-Assisted AnalysisLLM agents and tool use → Medical and scientific domain agents
Disentangling MLP Neuron Weights in Vocabulary SpaceMechanistic interpretability → Circuits and mechanistic analysis
Is CLIP Cross-Eyed? Revealing and Mitigating Center Bias in the CLIP FamilyMultimodal LLMs / vision-language models → Prompt learning and adaptation of CLIP
A Formal Security Framework for MCP-Based AI Agents: Threat Taxonomy, Verification Models, and Defense MechanismsLLM agents and tool use → Agent threats, guardrails and governance
FinReporting: An Agentic Workflow for Localized Reporting of Cross-Jurisdiction Financial DisclosuresLLM agents and tool use → Financial trading and forecasting agents
Beyond Compromise: Pareto-Lenient Consensus for Efficient Multi-Preference LLM AlignmentLLM alignment / RLHF / preference optimisation → LLM preference optimization methods
HumANDiff: Articulated Noise Diffusion for Motion-Consistent Human Video GenerationDiffusion models → Video generation
Does Pass Rate Tell the Whole Story? Evaluating Design Constraint Compliance in LLM-based Issue ResolutionLLM agents and tool use → Software engineering coding agents
Towards Trustworthy Report Generation: A Deep Research Agent with Progressive Confidence Estimation and CalibrationLLM agents and tool use → Web, search, and research agents
MARL-GPT: Foundation Model for Multi-Agent Reinforcement LearningLLM agents and tool use → Reinforcement learning for LLM agents
BOSCH: Black-Box Binary Optimization for Short-Context Attention-Head Selection in LLMsEfficient LLM inference and serving → KV cache optimization
Context-Value-Action Architecture for Value-Driven Large Language Model AgentsLLM agents and tool use → General LLM agent frameworks and benchmarks
Saliency-Guided Representation with Consistency Policy Learning for Visual Unsupervised Reinforcement LearningDeep reinforcement learning methods → RL for robot locomotion and manipulation
"I See What You Did There": Can Large Vision-Language Models Understand Multimodal Puns?Multimodal LLMs / vision-language models → MLLM training and visual instruction tuning
AICA-Bench: Holistically Examining the Capabilities of VLMs in Affective Image Content AnalysisMultimodal LLMs / vision-language models → Multimodal emotion understanding
HybridKV: Hybrid KV Cache Compression for Efficient Multimodal Large Language Model InferenceLLM reasoning → Multimodal reasoning +2
Addressing the Reasoning Gap: Mechanistic Circuit-Based Knowledge Editing in Large Language ModelsLLM reasoning → Knowledge- and tool-augmented reasoning +1
Understanding Performance Gap Between Parallel and Sequential Sampling in Large Reasoning ModelsLLM reasoning → Test-time compute and efficient thinking
Deep Researcher Agent: An Autonomous Framework for 24/7 Deep Learning Experimentation with Zero-Cost MonitoringLLM agents and tool use → Web, search, and research agents
Reading Between the Pixels: An Inscriptive Jailbreak Attack on Text-to-Image ModelsLLM safety and security → Vision-language and text-to-image safety
AgentGL: Towards Agentic Graph Learning with LLMs via Reinforcement LearningLLM agents and tool use → Web, search, and research agents
WikiSeeker: Rethinking the Role of Vision-Language Models in Knowledge-Based Visual Question AnsweringLLM reasoning → Multimodal reasoning +2
Adjustable Text-Guided Backdoor Attacks with Natural-Word Triggers on Multimodal Pretrained ModelsLLM safety and security → Backdoor and data poisoning attacks
Hierarchical Reinforcement Learning with Augmented Step-Level Transitions for LLM AgentsLLM agents and tool use → Reinforcement learning for LLM agents

All 498