Papers

New on arXiv

Wed 18 Feb 2026

323 papers

AgentFixer: From Failure Detection to Fix Recommendations in LLM Agentic SystemsLLM agents and tool use → Software engineering coding agents
Enhancing Financial Report Question-Answering: A Retrieval-Augmented Generation System with Reranking AnalysisRetrieval-augmented generation → Legal and financial RAG
Aligning Language Models from User InteractionsLLM alignment / RLHF / preference optimisation → LLM preference optimization methods
Task-Specific Knowledge Distillation via Intermediate ProbesKnowledge distillation → Knowledge distillation for LLMs
Know When You're Wrong: Aligning Confidence with Correctness for LLM Error DetectionLLM alignment / RLHF / preference optimisation → LLM preference optimization methods
RoboLayout: Differentiable 3D Scene Generation for Embodied AgentsLLM reasoning → Spatial and embodied reasoning +2
Auction-Based RIS Allocation With DRL: Controlling the Cost-Performance Trade-OffDeep reinforcement learning methods → RL for energy, networks and operations
Suppression or Deletion: A Restoration-Based Representation-Level Analysis of Machine UnlearningLLM safety and security → Machine unlearning
SIRUP: A diffusion-based virtual upmixer of steering vectors for highly-directive spatialization with first-order ambisonicsDiffusion models → Audio, speech and music
SAGE: Structure Aware Graph Expansion for Retrieval of Heterogeneous DataRetrieval-augmented generation → Knowledge-graph and structured-knowledge RAG
Greedy Multi-Path Block Verification for Faster Decoding in Speculative SamplingEfficient LLM inference and serving → Speculative decoding
Automating Agent Hijacking via Structural Template InjectionLLM agents and tool use → Agent threats, guardrails and governance
LLM4Cov: Execution-Aware Agentic Learning for High-coverage Testbench GenerationLLM agents and tool use → Software engineering coding agents
Exact Certification of Data-Poisoning Attacks Using Mixed-Integer ProgrammingLLM safety and security → Backdoor and data poisoning attacks
Mind the GAP: Text Safety Does Not Transfer to Tool-Call Safety in LLM AgentsLLM agents and tool use → Agent threats, guardrails and governance +1
DeepContext: Stateful Real-Time Detection of Multi-Turn Adversarial Intent Drift in LLMsLLM safety and security → Jailbreaking
Narrow Fine-Tuning Erodes Safety Alignment in Vision-Language AgentsMultimodal LLMs / vision-language models → MLLM jailbreaks and adversarial robustness +3
SparTa: Sparse Graphical Task Models from a Handful of DemonstrationsOffline RL and imitation learning from data → Robot manipulation imitation learning
AgentLAB: Benchmarking LLM Agents against Long-Horizon AttacksLLM agents and tool use → Agent threats, guardrails and governance
MALLVI: A Multi-Agent Framework for Integrated Generalized Robotics ManipulationLLM agents and tool use → Multimodal and embodied agents
OpenSage: Self-programming Agent Generation EngineLLM agents and tool use → Software engineering coding agents
AdaptOrch: Task-Adaptive Multi-Agent Orchestration in the Era of LLM Performance ConvergenceLLM agents and tool use → General LLM agent frameworks and benchmarks
DODO: Discrete OCR Diffusion ModelsDiffusion models → Image generation, text-to-image and editing
Overseeing Agents Without Constant Oversight: Challenges and OpportunitiesLLM agents and tool use → General LLM agent frameworks and benchmarks
Training Large Reasoning Models Efficiently via Progressive Thought EncodingLLM reasoning → Test-time compute and efficient thinking
NeST: Neuron Selective Tuning for LLM SafetyLLM alignment / RLHF / preference optimisation → Safety alignment and jailbreak attacks +1
IndicJR: A Judge-Free Benchmark of Jailbreak Robustness in South Asian LanguagesLLM alignment / RLHF / preference optimisation → Safety alignment and jailbreak attacks +1
Formal Mechanistic Interpretability: Automated Circuit Discovery with Provable GuaranteesMechanistic interpretability → Circuits and mechanistic analysis
Hybrid-Gym: Training Coding Agents to Generalize Across TasksLLM agents and tool use → Software engineering coding agents
Flow Map Language Models: One-step Language Modeling via Continuous DenoisingDiffusion models → Discrete and text diffusion
References Improve LLM Alignment in Non-Verifiable DomainsLLM alignment / RLHF / preference optimisation → Process reward models for reasoning
Escaping the Cognitive Well: Efficient Competition Math with Off-the-Shelf ModelsLLM reasoning → Mathematical reasoning
Better Think Thrice: Learning to Reason Causally with Double Counterfactual ConsistencyLLM reasoning → Causal reasoning in LLMs
Privacy-Aware Split Inference with Speculative Decoding for Large Language Models over Wide-Area NetworksEfficient LLM inference and serving → Early exit and edge inference
The Vulnerability of LLM Rankers to Prompt Injection AttacksLLM safety and security → Jailbreaking
PETS: A Principled Framework Towards Optimal Trajectory Allocation for Efficient Test-Time Self-ConsistencyLLM reasoning → Test-time compute and efficient thinking
DeepVision-103K: A Visually Diverse, Broad-Coverage, and Verifiable Mathematical Dataset for Multimodal ReasoningLLM reasoning → Multimodal reasoning +1
Formal Policy Enforcement for Real-World Agentic SystemsLLM agents and tool use → Agent threats, guardrails and governance
Saliency-Aware Multi-Route Thinking: Revisiting Vision-Language ReasoningLLM reasoning → Multimodal reasoning +1
Calibrate-Then-Act: Cost-Aware Exploration in LLM AgentsLLM agents and tool use → General LLM agent frameworks and benchmarks
Protecting the Undeleted in Machine UnlearningLLM safety and security → Machine unlearning
Retrieval-Augmented Foundation Models for Matched Molecular Pair Transformations to Recapitulate Medicinal Chemistry IntuitionRetrieval-augmented generation → Medical and biomedical RAG
Learning to unfold cloth: Scaling up world models to deformable object manipulationWorld models → Embodied and robot world models
SPARC: Scenario Planning and Reasoning for Automated C Unit Test GenerationLLM reasoning → Reasoning for agents, planning and code
Evaluating Collective Behaviour of Hundreds of LLM AgentsLLM agents and tool use → Social simulation and multi-agent debate
Align Once, Benefit Multilingually: Enforcing Multilingual Consistency for LLM Safety AlignmentLLM alignment / RLHF / preference optimisation → Safety alignment and jailbreak attacks +1
Agent Skill Framework: Perspectives on the Potential of Small to Medium Language Models in Industrial EnvironmentsLLM agents and tool use → Agent skill learning and evolution
Retrieval Augmented Generation of Literature-derived Polymer Knowledge: The Example of a Biodegradable Polymer Expert SystemRetrieval-augmented generation → Medical and biomedical RAG
Enhanced Diffusion Sampling: Efficient Rare Event Sampling and Free Energy Calculation with Diffusion ModelsDiffusion models → Flow matching, time series and science
Almost Sure Convergence of Differential Temporal Difference Learning for Average Reward Markov Decision ProcessesDeep reinforcement learning methods → RL theory and sample complexity
ColBERT-Zero: To Pre-train Or Not To Pre-train ColBERT modelsKnowledge distillation → General knowledge distillation methods
FlowPrefill: Decoupling Preemption from Prefill Scheduling Granularity to Mitigate Head-of-Line Blocking in LLM ServingEfficient LLM inference and serving → Serving systems and request scheduling
Quantifying Error Propagation and Model Collapse in Diffusion ModelsDiffusion models → Diffusion foundations and sampling +1
Sequential Membership Inference AttacksLLM safety and security → Membership inference attacks
Why Thinking Hurts: Diagnosing and Rectifying Linguistic Inertia in Large Language Models for RecommendationLLM reasoning → LLM reasoning for recommendation
DataJoint 2.0: A Computational Substrate for Agentic Scientific WorkflowsLLM agents and tool use → Web, search, and research agents
Steering diffusion models with quadratic rewards: a fine-grained analysisDiffusion models → Diffusion foundations and sampling
A Scalable Approach to Solving Simulation-Based Network Security GamesDeep reinforcement learning methods → Adversarial robustness and RL in finance
MerLean: An Agentic Framework for Autoformalization in Quantum ComputationLLM agents and tool use → Medical and scientific domain agents
RIDER: 3D RNA Inverse Design with Reinforcement Learning-Guided DiffusionDiffusion models → Flow matching, time series and science
Demonstration-Guided Observation Attacks on Black-Box Safe Reinforcement Learning Controllers for Robotic SystemsDeep reinforcement learning methods → Safe and constrained RL
Capacity-constrained demand response in smart grids using deep reinforcement learningDeep reinforcement learning methods → RL for energy, networks and operations
Reinforcement Learning for Parameterized Quantum State Preparation: A Comparative StudyDeep reinforcement learning methods → Quantum reinforcement learning
Recursive language models for jailbreak detection: a procedural defense for tool-augmented agentsLLM agents and tool use → Agent threats, guardrails and governance +1
Framework of Thoughts: A Foundation Framework for Dynamic and Optimized Reasoning based on Chains, Trees, and GraphsLLM reasoning → Prompting and chain-of-thought reasoning
Fast and Scalable Analytical DiffusionDiffusion models → Diffusion foundations and sampling
MMA: Multimodal Memory AgentLLM agents and tool use → Agent memory systems
Team of Thoughts: Efficient Test-time Scaling of Agentic Systems through Orchestrated Tool CallingLLM reasoning → Reasoning for agents, planning and code +1
Certifying Hamilton-Jacobi Reachability Learned via Reinforcement LearningDeep reinforcement learning methods → RL theory and sample complexity
Visual Self-Refine: A Pixel-Guided Paradigm for Accurate Chart ParsingMultimodal LLMs / vision-language models → Chart understanding and anomaly detection
RoboGene: Boosting VLA Pre-training via Diversity-Driven Agentic Framework for Real-World Task GenerationLLM agents and tool use → Multimodal and embodied agents
Intra-Fairness Dynamics: The Bias Spillover Effect in Targeted LLM AlignmentLLM alignment / RLHF / preference optimisation → Human values and AI feedback alignment
TabAgent: A Framework for Replacing Agentic Generative Components with Tabular-Textual ClassifiersLLM agents and tool use → General LLM agent frameworks and benchmarks
Verifiable Semantics for Agent-to-Agent CommunicationLLM agents and tool use → General LLM agent frameworks and benchmarks
ReMoRa: Multimodal Large Language Model based on Refined Motion Representation for Long-Video UnderstandingMultimodal LLMs / vision-language models → Video understanding with MLLMs
Easy Data Unlearning BenchLLM safety and security → Machine unlearning
Label-Consistent Data Generation for Aspect-Based Sentiment Analysis Using LLM AgentsLLM agents and tool use → General LLM agent frameworks and benchmarks
Dual-Quadruped Collaborative Transportation in Narrow Environments via Safe Reinforcement LearningDeep reinforcement learning methods → Safe and constrained RL
Helpful to a Fault: Measuring Illicit Assistance in Multi-Turn, Multilingual LLM AgentsLLM agents and tool use → Agent threats, guardrails and governance +1
Spatial Audio Question Answering and Reasoning on Dynamic Source MovementsLLM reasoning → Audio-language model reasoning
MemoryArena: Benchmarking Agent Memory in Interdependent Multi-Session Agentic TasksLLM agents and tool use → Agent memory systems
Regret and Sample Complexity of Online Q-Learning via Concentration of Stochastic Approximation with Time-Inhomogeneous Markov ChainsDeep reinforcement learning methods → RL theory and sample complexity
Toward Scalable Verifiable Reward: Proxy State-Based Evaluation for Multi-turn Tool-Calling LLM AgentsLLM agents and tool use → General LLM agent frameworks and benchmarks
Factored Latent Action World ModelsWorld models → Embodied and robot world models
Training-Free Adaptation of Diffusion Models via Doob's h-TransformDiffusion models → Diffusion foundations and sampling +1
ModalImmune: Immunity Driven Unlearning via Self Destructive TrainingLLM safety and security → Machine unlearning
Deep TPC: Temporal-Prior Conditioning for Time Series ForecastingLLM reasoning → Time series reasoning
World Model Failure Classification and Anomaly Detection for Autonomous InspectionWorld models → Latent dynamics models and model-based RL
EnterpriseBench Corecraft: Training Generalizable Agents on High-Fidelity RL EnvironmentsLLM agents and tool use → General LLM agent frameworks and benchmarks
Learning Personalized Agents from Human FeedbackLLM alignment / RLHF / preference optimisation → Reward modeling and RLHF
Discrete Stochastic Localization for Non-autoregressive GenerationDiffusion models → Diffusion foundations and sampling
HiPER: Hierarchical Reinforcement Learning with Explicit Credit Assignment for Large Language Model AgentsLLM agents and tool use → Reinforcement learning for LLM agents
Balancing Faithfulness and Performance in Reasoning via Multi-Listener Soft ExecutionLLM reasoning → Prompting and chain-of-thought reasoning
IRIS: Intent Resolution via Inference-time Saccades for Open-Ended VQA in Large Vision-Language ModelsMultimodal LLMs / vision-language models → VLM benchmarks and visual reasoning
Retrieval Collapses When AI Pollutes the WebRetrieval-augmented generation → Privacy and security of RAG
CHAI: CacHe Attention Inference for text2videoDiffusion models → Video generation
Empirical Cumulative Distribution Function Clustering for LLM-based Agent System AnalysisLLM agents and tool use → General LLM agent frameworks and benchmarks
GPSBench: Do Large Language Models Understand GPS Coordinates?LLM reasoning → Spatial and embodied reasoning
Neuro-Symbolic Strong-AI Robots with Closed Knowledge Assumption: Learning and Deductions
LETGAMES: An LLM-Powered Gamified Approach to Cognitive Training for Patients with Cognitive Impairment

All 323