Papers

New on arXiv

Mon 17 Aug 2026

439 papers

Financial Tail Risk Beyond Lipschitz Continuity via Semi-Discrete Optimal TransportDiffusion models → Flow matching, time series and science
Beyond Unsafe Detection: Counterfactually Anchored Evidence Attribution for Multi-Turn LLM Safety FailuresLLM agents and tool use → Agent threats, guardrails and governance +1
Differentiable Fuzzy Inference Layer: A Monotone, Compositional Ordinal Reasoning Head for Large Language ModelsLLM reasoning → Reasoning benchmarks and evaluation
Certified Mechanistic Interpretability: Lifting Single-Input Findings to Bounded NeighbourhoodsMechanistic interpretability → Circuits and mechanistic analysis
A Shared Learning Rate Is Not a Neutral Control in Selective On-Policy DistillationKnowledge distillation → On-policy and self-distillation for LLMs
AI Revealed PreferencesLLM alignment / RLHF / preference optimisation → Human values and AI feedback alignment
Boosting Knowledge-based Visual Question Answering with Structured Context ReasoningLLM reasoning → Multimodal reasoning
Evaluating Multimodal Narrative Understanding of Popular Hollywood FilmsMultimodal LLMs / vision-language models → Video understanding with MLLMs
Looped Language Models Improve Compositional Tool CallingLLM agents and tool use → Tool learning and function calling
Teach and Grow: An Agent-Centered Architecture for General Robot LearningWorld models → Embodied and robot world models +2
Which Source Wins? Task-Dependent Reliance in Vision-Language ModelsMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning
Expressivity In Multimodal Contrastive LearningMultimodal LLMs / vision-language models → Prompt learning and adaptation of CLIP
Fool's Gold: Defensive Deception Against Safety-Removal Attacks on Open-Weight ModelsLLM alignment / RLHF / preference optimisation → Safety alignment and jailbreak attacks +1
Graphectory Viewer: A Tool for Process-Centric Analysis of Agentic Software TrajectoriesLLM agents and tool use → Software engineering coding agents
Token Optimization and Context Window Management in Multi-Agent AI WorkflowsLLM agents and tool use → Agent memory systems
Reinforcement Learning as (Discrete) Potential TheoryDeep reinforcement learning methods → RL theory and sample complexity
Task Specialization Fine-Tuning for Contextual Reinforcement LearningDeep reinforcement learning methods → RL training of language and agent models
Grounding AI Agents in Contracts: An Empirical Evaluation of Spec-Driven Test GenerationLLM agents and tool use → Software engineering coding agents
Automating Parent Selection Configuration in Genetic Programming with Agentic AILLM reasoning → Reasoning for agents, planning and code
Polaris: Learning to Generate Table Descriptions from Retrieval FeedbackLLM alignment / RLHF / preference optimisation → Preference alignment for recommenders and RAG
Synthesizing Feature Extractors: An Agentic Approach for Algorithm SelectionLLM agents and tool use → General LLM agent frameworks and benchmarks
Can LLMs Reason in a Legally Meaningful Manner? A Small-scale Study on European Court of Human Rights CasesLLM reasoning → Reasoning for agents, planning and code
Q-Learning With World ModelsWorld models → Latent dynamics models and model-based RL +2
Reasoning Reduces the Influence of Poisoned Context in RAGLLM reasoning → Reasoning for agents, planning and code +1
Lymphocyte Mimicry Correction via Region-Level Tissue Reasoning and Unbalanced Optimal TransportLLM reasoning → Clinical and biomedical applications
Authorization Before Context: A Model-Neutral Audience Boundary Against Cross-Audience Memory Leakage in Agentic SystemsLLM agents and tool use → Agent memory systems
PDDL-ART: Autonomous Symbolic Abstraction From Demonstration For Long-Horizon Robotic Manipulation Using Vision-Language ModelsMultimodal LLMs / vision-language models → Embodied and 3D vision-language models
PROBE: Manipulation-Grounded Visual Question Answering with VLM AgentsLLM reasoning → Spatial and embodied reasoning +2
Physics-informed Reinforcement Learning for Stochastic Reach-Avoid AnalysisDeep reinforcement learning methods → RL theory and sample complexity
Appearing Legitimate is Not Enough: Interrogating Synthetic Agents in Representational Processes through a Participatory Design LensLLM agents and tool use → Social simulation and multi-agent debate
Inference-Time Attention Steering for Vision-Language-Action Driving ModelsVision-language-action models → VLA for autonomous driving +1
Uncertainty-Aware Decision Making in Multimodal Large Language ModelsMultimodal LLMs / vision-language models → MLLM training and visual instruction tuning
Foundation Agents Meet Agentic Deep Research: Evidence-Grounded Clinical Code ForecastingLLM agents and tool use → Medical and scientific domain agents
DiSCO: Defending text-to-image generation through distribution-guided contrastive prompt optimizationLLM safety and security → Vision-language and text-to-image safety
Wasted large language models: A life cycle thinking approachLLM reasoning → Reasoning for agents, planning and code
Lambda-Hold Control: Human-Like Movement Emerges from a Minimal Task Reward in Predictive Musculoskeletal SimulationDeep reinforcement learning methods → RL for robot locomotion and manipulation
LadderTeam: Dual-Agent Laddering Elicitation FrameworkLLM agents and tool use → Software engineering coding agents
SkillEffect: Checked Lowering for Memory-Bounded Agent ToolsLLM agents and tool use → Agent skill learning and evolution
VLCP: Vision Language Control Policy Closed-Loop Code Replanning for Robot ManipulationMultimodal LLMs / vision-language models → Embodied and 3D vision-language models
Don't Drop the BATON: Long-Horizon Robot Manipulation via Agentic Subtask Exploration and Transition-aware MemoryLLM agents and tool use → Multimodal and embodied agents
Q-based Variational Inverse Reinforcement LearningOffline RL and imitation learning from data → Inverse reinforcement learning +1
An Empirical Study of Training Pixel-Space Text-to-Image Diffusion ModelsDiffusion models → Diffusion foundations and sampling
τ0-VLA: a Hierarchical Robot Foundation Model with World-Model-Guided Test-Time ComputationWorld models → Embodied and robot world models
HarnessEval-W: Agentifying the Evaluation of Visual WorldsWorld models → World models for language agents
Security of Foundation-Model-Powered Embodied Agents: Attack Surfaces, Attacks, Defenses, and EvaluationLLM agents and tool use → Agent threats, guardrails and governance
HAF: Adapting Generalist VLAs to Humanoid Whole-Body Loco-manipulation via Hierarchical Action Flow and Spectral Latent RLVision-language-action models → Tactile, force-aware and humanoid VLA +1
Policy Iteration with Human Feedback: Bringing Post-Training RL to In-context LearningLLM alignment / RLHF / preference optimisation → Reward modeling and RLHF
CaliBench: Are the Stochastic Dynamics of Video World Models Physically Calibrated?World models → Video-generation world models and simulators
Unlocking the Potential of Image Editing via Concept Scaling and Dense SupervisionDiffusion models → Image generation, text-to-image and editing
Breaking Planner Integrity Boundary: Enviroment State-Text Injection Attack on LLM-Driven Embodied AgentsLLM agents and tool use → Agent threats, guardrails and governance
Diagnosing Dense Same-Class Attribute Misbinding in Large Vision-Language ModelsMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning
When Agents Coordinate: Measuring Coordination in Multi-Agent AI CodingLLM agents and tool use → General LLM agent frameworks and benchmarks
Neurosymbolic Embodied AgentsMultimodal LLMs / vision-language models → Embodied and 3D vision-language models +1
PixRestore: Unified Image Restoration via Pixel Diffusion TransformerDiffusion models → Inverse problems and image restoration
Steering the Flow: Inverting Face Recognition Models via Gradient-Guided Flow MatchingDiffusion models → Robustness, privacy and adversarial aspects
Revisiting Classifier-Free Guidance Methods in Latent Diffusion ModelsDiffusion models → Diffusion foundations and sampling
GRIP: Grounded Reasoning via Information-Restricted PremisesRetrieval-augmented generation → RAG robustness and adaptive retrieval
Would this change your answer? Evaluating Explanations of LLM Behavior In The Wild with Counterfactual ExperimentsLLM reasoning → Causal reasoning in LLMs
TDD-Agent: Test-Driven Reasoning for Code GenerationLLM reasoning → Reasoning for agents, planning and code +1
GenRouter: Unified Workflow Routing for Agentic Image GenerationLLM agents and tool use → Multimodal and embodied agents
MatchingPolicy: Correspondence-Aware Policy Enables Cross-Object In-Context LearningOffline RL and imitation learning from data → Robot manipulation imitation learning
The Ethical Decision Head: Operationalizing Normative Ethics in Autonomous Vehicles via Reinforcement Learning from Human FeedbackLLM alignment / RLHF / preference optimisation → Human values and AI feedback alignment +1
Learning to Unlearn: Machine Unlearning via Learning the Unlearning BehaviorsLLM safety and security → Machine unlearning
FabriMAE I Trust Myself? Self-Evaluating VLA Action Generation with Markov Attention EntropyVision-language-action models → VLA architectures and generalization
AnchorScore: A CLIP-Based Diagnostic of MLLM Annotation DifficultyMultimodal LLMs / vision-language models → MLLM training and visual instruction tuning
Chronocooked: A Benchmark for Interval Timing in Reinforcement Learning AgentsDeep reinforcement learning methods → Deep RL algorithms and representations
Orbit-Planner: Towards Latent World Models for On-Orbit Obstacle Avoidance of Satellite AgentsWorld models → World models for navigation
Every Coin Has Two Sides: On the Dual Nature of Generalization in On-Policy Distillation of Large Language ModelsKnowledge distillation → On-policy and self-distillation for LLMs
PDDLCoder: Agentic PDDL Generation for LLM-Assisted Symbolic PlanningLLM agents and tool use → General LLM agent frameworks and benchmarks
The Working Set of a Coding Agent: Coherence Debt in Repository-Scale TasksLLM agents and tool use → Software engineering coding agents
Hypergraph-based Multimodal Retrieval-Augmented Generation with Incremental RefinementRetrieval-augmented generation → Multimodal RAG
Cost Scales with Change, Not Corpus Size: Incrementally Maintaining an Evolving Semantic SubstrateRetrieval-augmented generation → Generation and context integration +1
Palmyra x6 Technical Report: An Agentic, Tool-Use Model Post-Trained via Anchored Supervised Fine-TuningLLM agents and tool use → General LLM agent frameworks and benchmarks
SQuad: Sub-Quadratic Attention Distillation for Efficient Video GenerationDiffusion models → Video generation
BabelSteering: Multilingual Safety Alignment via English Steering VectorsLLM alignment / RLHF / preference optimisation → Safety alignment and jailbreak attacks +2
Ask, Condition or Abstain: Reinforcement Learning for Missing-Premise ReasoningLLM reasoning → Reinforcement learning for reasoning
STAGE: Controlled Objective Admission for Multi-Preference LLM AlignmentLLM alignment / RLHF / preference optimisation → LLM preference optimization methods
What to Remember, What to Reveal: Privacy-Aware Memory for Conversational AgentsLLM agents and tool use → Agent threats, guardrails and governance
Listen, Reason, and Segment: Aligning LALMs with Editorial Judgment for Media ChapterizationLLM reasoning → Audio-language model reasoning
DSPrompt: Dynamic Soft Prompt Defense Against M-RAG CorruptionRetrieval-augmented generation → Privacy and security of RAG +1
FLEET: Token-Based Feature Extraction for Event Camera-based Reinforcement LearningDeep reinforcement learning methods → RL for robot locomotion and manipulation
When Context Misleads: Intent-Guided Decoding for Robust Retrieval-Augmented GenerationRetrieval-augmented generation → RAG robustness and adaptive retrieval
MLLM-Guided Semantic Correction for Text-to-Video GenerationDiffusion models → Video generation
NebulaVLA: A Dual-Frequency Vision-Language-Action Model With Guide Action for Robotic ManipulationVision-language-action models → Generalist VLA models and benchmarks
RISE: Roadside Infrastructure Sequence Understanding across 3D Tracking and Structured Vision-Language ReasoningLLM reasoning → Chain-of-thought reasoning for driving
Pallas: A Proactive KV Cache Migration Framework for LLM Inference in AI-RANEfficient LLM inference and serving → KV cache optimization
JailbreakSkill: Scaling Automated Red-Teaming with Reusable and Ever-Evolving SkillsLLM safety and security → Prompt injection and agent security
HaReCAP: Habitual-action Grounding for Recursive Large Language Model AgentsLLM agents and tool use → General LLM agent frameworks and benchmarks
Time to Reason: Scalable Neurosymbolic Learning for LTLf via Fuzzy SemanticsLLM reasoning → Reasoning benchmarks and evaluation
Stable Multi-Step Rollouts via Uncertainty-Guided Hybrid DynamicsDeep reinforcement learning methods → RL theory and sample complexity
ParaTempo: Efficient Parallel Reasoning via Temporal ConfidenceLLM reasoning → Test-time compute and efficient thinking
Joint Flow Matching Enables Continuous Dose-Conditioned Cell MorphingDiffusion models → Flow matching, time series and science
PertMind: Eliciting Emergent Biological Reasoning in LLM via Reinforcement Learning on Cellular Perturbation DataLLM reasoning → Clinical and biomedical applications
D2-ScaleAgent: Dual-Dimensional Scaling for Long Document UnderstandingRetrieval-augmented generation → Agentic RAG, agents and memory
Towards Risk-free AI Agent DeploymentLLM agents and tool use → Software engineering coding agents
TRACE-CASH: Trial-History-Conditioned Reinforcement Learning for Adaptive Configuration Exploration in Time-Series CASHDeep reinforcement learning methods → RL training of language and agent models
A Policy Algebra for Trust-Preserving Agentic AI ExecutionLLM agents and tool use → Agent threats, guardrails and governance
Think Inside the Chunk: RegulaRAG for Regulation-Compliant Scenario Generation using LLMs: A Case Study of UN Regulation No. 152LLM reasoning → Knowledge- and tool-augmented reasoning +1
Security Assessment of DeepSeek Harness with A.I.G: Evaluating Resistance to Indirect Prompt InjectionLLM safety and security → Prompt injection and agent security
Mint-Agent: Introducing Finance-Native Agentic Foundation ModelsLLM agents and tool use → Financial trading and forecasting agents

All 439