Papers

New on arXiv

Fri 9 Jan 2026

286 papers

A Framework for Low-Latency, LLM-driven Multimodal Interaction on the Pepper RobotLLM agents and tool use → Multimodal and embodied agents
UniFinEval: Towards Unified Evaluation of Financial Multimodal Models across Text, Images and VideosLLM reasoning → Reasoning for agents, planning and code +1
Demystifying Multi-Agent Debate: The Role of Confidence and DiversityLLM reasoning → Reasoning for agents, planning and code +1
Mixture-of-Experts as Soft Clustering: A Dual Jacobian-PCA Spectral Geometry PerspectiveMixture-of-experts → Expert routing strategies
CTHA: Constrained Temporal Hierarchical Architecture for Stable Multi-Agent LLM SystemsLLM agents and tool use → General LLM agent frameworks and benchmarks
Bias Detection and Rotation-Robustness Mitigation in Vision-Language Models and Generative Image ModelsMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning
FinVault: Benchmarking Financial Agent Safety in Execution-Grounded EnvironmentsLLM agents and tool use → Financial trading and forecasting agents +1
Monkey Jump : MoE-Style PEFT for Efficient Multi-Task LearningMixture-of-experts → Mixture of LoRA experts
CARD: Cluster-level Adaptation with Reward-guided Decoding for Personalized Text GenerationLLM alignment / RLHF / preference optimisation → Human values and AI feedback alignment
Circuit Mechanisms for Spatial Relation Generation in Diffusion TransformersDiffusion models → Image generation, text-to-image and editing
C-World: A Computer Use Agent Environment CreatorLLM agents and tool use → General LLM agent frameworks and benchmarks
Why LoRA Fails to Forget: Regularized Low-Rank Adaptation Against Backdoors in Language ModelsLLM safety and security → Backdoor and data poisoning attacks
How well can off-the-shelf LLMs elucidate molecular structures from mass spectra using chain-of-thought reasoning?LLM reasoning → Clinical and biomedical applications
AIConfigurator: Lightning-Fast Configuration Optimization for Multi-Framework LLM ServingEfficient LLM inference and serving → Serving systems and request scheduling
Amory: Building Coherent Narrative-Driven Agent Memory through Agentic ReasoningLLM agents and tool use → Agent memory systems
Automated QoR improvement in OpenROAD with coding agentsLLM agents and tool use → Software engineering coding agents
Agentic AI Microservice Framework for Deepfake and Document Fraud Detection in KYC PipelinesLLM agents and tool use → Agent threats, guardrails and governance
An Intelligent AI glasses System with Multi-Agent Architecture for Real-Time Voice Processing and Task ExecutionLLM agents and tool use → General LLM agent frameworks and benchmarks
PCoKG: Personality-aware Commonsense Reasoning with DebateLLM reasoning → Reasoning benchmarks and evaluation
Synthetic FMCW Radar Range Azimuth Maps Augmentation with Generative Diffusion ModelDiffusion models → 3D, point cloud and scene generation
Ground What You See: Hallucination-Resistant MLLMs via Caption Feedback, Diversity-Aware Sampling, and Conflict RegularizationMultimodal LLMs / vision-language models → Hallucination mitigation in LVLMs
AdaFuse: Adaptive Ensemble Decoding with Test-Time Scaling for LLMsLLM reasoning → Test-time compute and efficient thinking
Chaining the Evidence: Robust Reinforcement Learning for Deep Search Agents with Citation-Aware Rubric RewardsLLM reasoning → Knowledge- and tool-augmented reasoning +2
Don't Break the Cache: An Evaluation of Prompt Caching for Long-Horizon Agentic TasksLLM agents and tool use → Software engineering coding agents
The Molecular Structure of Thought: Mapping the Topology of Long Chain-of-Thought ReasoningLLM reasoning → Prompting and chain-of-thought reasoning
VideoAR: Autoregressive Video Generation via Next-Frame & Scale PredictionDiffusion models → Video generation
Distilling Feedback into Memory-as-a-ToolLLM reasoning → Test-time compute and efficient thinking
Context-Aware Decoding for Faithful Vision-Language GenerationLLM reasoning → Hallucination detection and mitigation +1
Can We Predict Before Executing Machine Learning Agents?LLM agents and tool use → Web, search, and research agents
Agentic LLMs as Powerful Deanonymizers: Re-identification of Participants in the Anthropic Interviewer DatasetLLM agents and tool use → Social simulation and multi-agent debate
Distilling Lightweight Domain Experts from Large ML Models by Identifying Relevant SubspacesKnowledge distillation → General knowledge distillation methods
TowerMind: A Tower Defence Game Learning Environment and Benchmark for LLM as AgentsLLM agents and tool use → Reinforcement learning for LLM agents
StackPlanner: A Centralized Hierarchical Multi-Agent System with Task-Experience Memory ManagementLLM agents and tool use → Agent memory systems
An Empirical Study on Preference Tuning Generalization and Diversity Under Domain ShiftLLM alignment / RLHF / preference optimisation → LLM preference optimization methods
iReasoner: Trajectory-Aware Intrinsic Reasoning Supervision for Self-Evolving Large Multimodal ModelsLLM reasoning → Multimodal reasoning +1
IIB-LPO: Latent Policy Optimization via Iterative Information BottleneckLLM reasoning → Reinforcement learning for reasoning
Sequential Bayesian Optimal Experimental Design in Infinite Dimensions via Policy Gradient Reinforcement LearningDeep reinforcement learning methods → RL theory and sample complexity
FACTUM: Mechanistic Detection of Citation Hallucination in Long-Form RAGRetrieval-augmented generation → RAG robustness and adaptive retrieval
CLewR: Curriculum Learning with Restarts for Machine Translation Preference LearningLLM alignment / RLHF / preference optimisation → LLM preference optimization methods
Kidney Cancer Detection Using 3D-Based Latent Diffusion ModelsDiffusion models → Medical imaging and segmentation
Goal Force: Teaching Video Models To Accomplish Physics-Conditioned GoalsWorld models → Latent predictive world models
Intelligent Singularity Avoidance in UR10 Robotic Arm Path Planning Using Hybrid Fuzzy Logic and Reinforcement LearningDeep reinforcement learning methods → RL for robot locomotion and manipulation
Boosting Latent Diffusion Models via Disentangled Representation AlignmentDiffusion models → Diffusion foundations and sampling
EnvScaler: Scaling Tool-Interactive Environments for LLM Agent via Programmatic SynthesisLLM agents and tool use → Tool learning and function calling
Modular Autonomy with Conversational Interaction: An LLM-driven Framework for Decision Making in Autonomous DrivingEnd-to-end autonomous driving → Large language models for driving
From Off-Policy to On-Policy: Enhancing GUI Agents via Bi-level Expert-to-Policy AssimilationLLM agents and tool use → GUI and mobile agents
EET: Experience-Driven Early Termination for Cost-Efficient Software Engineering AgentsLLM agents and tool use → Software engineering coding agents
VIGIL: Defending LLM Agents Against Tool Stream Injection via Verify-Before-CommitLLM agents and tool use → Agent threats, guardrails and governance +1
DynaDebate: Breaking Homogeneity in Multi-Agent Debate with Dynamic Path GenerationLLM agents and tool use → Social simulation and multi-agent debate
The Echo Chamber Multi-Turn LLM JailbreakLLM safety and security → Jailbreaking
PII-VisBench: Evaluating Personally Identifiable Information Safety in Vision Language Models Along a Continuum of VisibilityMultimodal LLMs / vision-language models → MLLM jailbreaks and adversarial robustness
TAGRPO: Boosting GRPO on Image-to-Video Generation with Direct Trajectory AlignmentLLM reasoning → Reinforcement learning for reasoning +1
Overcoming Joint Intractability with Lossless Hierarchical Speculative DecodingEfficient LLM inference and serving → Speculative decoding
Rotate Your Character: Revisiting Video Diffusion Models for High-Quality 3D Character GenerationDiffusion models → 3D, point cloud and scene generation
Logic-Parametric Neuro-Symbolic NLI: Controlling Logical Formalisms for Verifiable LLM ReasoningLLM reasoning → Reasoning benchmarks and evaluation
Circular Reasoning: Understanding Self-Reinforcing Loops in Large Reasoning ModelsLLM reasoning → Test-time compute and efficient thinking
SketchVL: Policy Optimization via Fine-Grained Credit Assignment for Chart Understanding and MoreLLM reasoning → Multimodal reasoning
Do Sparse Autoencoders Identify Reasoning Features in Language Models?LLM reasoning → Reasoning benchmarks and evaluation +1
CHDP: Cooperative Hybrid Diffusion Policies for Reinforcement Learning in Parameterized Action SpaceDiffusion models → Diffusion and flow policies for robotics +1
HAG: Hierarchical Demographic Tree-based Agent Generation for Topic-Adaptive SimulationLLM agents and tool use → Social simulation and multi-agent debate
EvoQRE: Modeling Bounded Rationality in Safety-Critical Traffic Simulation via Evolutionary Quantal Response EquilibriumOffline RL and imitation learning from data → Imitation and offline RL for driving
Multilingual Amnesia: On the Transferability of Unlearning in Multilingual LLMsLLM safety and security → Machine unlearning
SGDrive: Scene-to-Goal Hierarchical World Cognition for Autonomous DrivingMultimodal LLMs / vision-language models → Driving VLMs and open-vocabulary detection +1
Continual Pretraining on Encrypted Synthetic Data for Privacy-Preserving LLMsSynthetic data and self-improvement for LLMs → Privacy-preserving tabular data synthesis
GIFT: Games as Informal Training for Generalizable LLMsLLM reasoning → Reinforcement learning for reasoning
Dual-Phase LLM Reasoning: Self-Evolved Mathematical FrameworksLLM reasoning → Test-time compute and efficient thinking +1
FLARE: Learning Future-Aware Latent Representations from Vision-Language Models for Autonomous DrivingMultimodal LLMs / vision-language models → Driving VLMs and open-vocabulary detection +1
Data Augmented Pipeline for Legal Information Extraction and ReasoningLLM reasoning → Reasoning for agents, planning and code
Orchestrating Tokens and Sequences: Dynamic Hybrid Policy Optimization for RLVRLLM reasoning → Reinforcement learning for reasoning
Conformity Dynamics in LLM Multi-Agent Systems: The Roles of Topology and Self-Social WeightingLLM agents and tool use → Social simulation and multi-agent debate
SceneAlign: Aligning Multimodal Reasoning to Scene Graphs in Complex Visual ScenesLLM reasoning → Multimodal reasoning +2
PaCoRe: Learning to Scale Test-Time Compute with Parallel Coordinated ReasoningLLM reasoning → Test-time compute and efficient thinking
Autonomous Discovery of the Ising Model's Critical Parameters with Reinforcement LearningDeep reinforcement learning methods → RL theory and sample complexity
Crisis-Bench: Benchmarking Strategic Ambiguity and Reputation Management in Large Language ModelsLLM alignment / RLHF / preference optimisation → Human values and AI feedback alignment
WildSci: Advancing Scientific Reasoning from In-the-Wild LiteratureLLM reasoning → Knowledge- and tool-augmented reasoning
ReasonAny: Incorporating Reasoning Capability to Any Model via Simple and Effective Model MergingLLM reasoning → Test-time compute and efficient thinking
Efficient Temporal-aware Matryoshka Adaptation for Temporal Information RetrievalLLM reasoning → Knowledge- and tool-augmented reasoning +1
VIB-Probe: Detecting and Mitigating Hallucinations in Vision-Language Models via Variational Information BottleneckMultimodal LLMs / vision-language models → Hallucination mitigation in LVLMs
Closing the Modality Reasoning Gap for Speech Large Language ModelsLLM reasoning → Audio-language model reasoning
LIDL: LLM Integration Defect Localization via Knowledge Graph-Enhanced Multi-Agent AnalysisLLM agents and tool use → Software engineering coding agents
Scalable Heterogeneous Graph Learning via Heterogeneous-aware Orthogonal Prototype ExpertsMixture-of-experts → MoE for time series and graphs
Learning specifications for reactive synthesis with safety constraintsOffline RL and imitation learning from data → Robot learning from demonstration
DeMa: Dual-Path Delay-Aware Mamba for Efficient Multivariate Time Series AnalysisState-space sequence models (S4, Mamba) → Time series forecasting with SSMs
Double: Breaking the Acceleration Limit via Double Retrieval Speculative ParallelismEfficient LLM inference and serving → Speculative decoding
CHisAgent: A Multi-Agent Framework for Event Taxonomy Construction in Ancient Chinese Cultural SystemsLLM agents and tool use → Multimodal and embodied agents
Memory Poisoning Attack and Defense on Memory Based LLM-AgentsLLM agents and tool use → Agent threats, guardrails and governance
MMViR: A Multi-Modal and Multi-Granularity Representation for Long-range Video UnderstandingMultimodal LLMs / vision-language models → Video understanding with MLLMs
MaxCode: A Max-Reward Reinforcement Learning Framework for Automated Code OptimizationDeep reinforcement learning methods → RL training of language and agent models
STELP: Secure Transpilation and Execution of LLM-Generated ProgramsLLM agents and tool use → Agent threats, guardrails and governance
Jailbreaking Large Language Models through Iterative Tool-Disguised Attacks via Reinforcement LearningLLM safety and security → Jailbreaking
PRISMA: Reinforcement Learning Guided Two-Stage Policy Optimization in Multi-Agent Architecture for Open-Domain Multi-Hop Question AnsweringRetrieval-augmented generation → Agentic RAG, agents and memory
Do LLMs Need Inherent Reasoning Before Reinforcement Learning? A Study in Korean Self-CorrectionLLM reasoning → Reasoning benchmarks and evaluation
ART: Adaptive Reasoning Trees for Explainable Claim VerificationLLM reasoning → Reasoning benchmarks and evaluation
RingSQL: Schema-Independent Synthetic Data Generation for Text-to-SQL Reinforcement LearningSynthetic data and self-improvement for LLMs → Synthetic reasoning data and bootstrapping +1
TAPM-Net: Trajectory-Aware Perturbation Modeling for Infrared Small Target DetectionState-space sequence models (S4, Mamba) → Point clouds, detection and remote sensing
Knowledge-Driven Multi-Turn Jailbreaking on Large Language ModelsLLM safety and security → Jailbreaking
Tracing Moral Foundations in Large Language ModelsMechanistic interpretability → Probing deception, bias and beliefs
OVT-MLCS: An Online Visual Tool for MLCS Mining from Long or Big Sequences
EviSnap: Faithful Evidence-Cited Explanations for Cold-Start Cross-Domain Recommendation
LLM-Augmented Knowledge Base Construction For Root Cause Analysis

All 286