Papers

New on arXiv

Tue 6 Jan 2026

357 papers

M-RAG: Semantic Key-Value Indexing for Retrieval-Augmented GenerationRetrieval-augmented generation → RAG pipeline design and evaluation
Generalizable Prompt Tuning for Audio-Language Models via Semantic ExpansionMultimodal LLMs / vision-language models → Prompt learning and adaptation of CLIP
Simulating Complex Multi-Turn Tool Calling Interactions in Stateless Execution EnvironmentsLLM agents and tool use → Tool learning and function calling
AfriEconQA: A Benchmark for Quantitative and Temporal Reasoning over World Bank Economic ReportsLLM reasoning → Reasoning benchmarks and evaluation +1
Enhancing Model Context Protocol (MCP) with Context-Aware Server CollaborationLLM agents and tool use → Tool learning and function calling
A Preliminary Agentic Framework for Matrix DeflationLLM agents and tool use → Multimodal and embodied agents
Forget Many, Forget Right: Scalable and Precise Concept Unlearning in Diffusion ModelsDiffusion models → Robustness, privacy and adversarial aspects +1
Student Guides Teacher: Weak-to-Strong Inference via Spectral Orthogonal ExplorationLLM reasoning → Mathematical reasoning
ECLIPTICA -- A Framework for Switchable LLM Alignment via CITA - Contrastive Instruction-Tuned AlignmentLLM alignment / RLHF / preference optimisation → LLM preference optimization methods
Channel Knowledge Map Construction via Guided Flow MatchingDiffusion models → Diffusion for compression and wireless
Scaling Trends for Multi-Hop Contextual Reasoning in Mid-Scale Language ModelsLLM reasoning → Knowledge- and tool-augmented reasoning +2
EpiQAL: Benchmarking Large Language Models in Epidemiological Question Answering and ReasoningLLM reasoning → Clinical and biomedical applications
Understanding Reward Hacking in Text-to-Image Reinforcement LearningLLM alignment / RLHF / preference optimisation → Diffusion and multimodal model alignment +1
ThinkRL-Edit: Thinking in Reinforcement Learning for Reasoning-Centric Image EditingLLM reasoning → Multimodal reasoning
Prompting Underestimates LLM Capability for Time Series ClassificationMechanistic interpretability → Interpreting time-series and weather models
FROST-Drive: Scalable and Efficient End-to-End Driving with a Frozen Vision EncoderEnd-to-end autonomous driving → End-to-end system architecture
Automated Feedback Generation for Undergraduate Mathematics: Development and Evaluation of an AI Teaching AssistantLLM reasoning → Mathematical reasoning
Grading Scale Impact on LLM-as-a-Judge: Human-LLM Alignment Is Highest on 0-5 Grading ScaleLLM alignment / RLHF / preference optimisation → Human values and AI feedback alignment
Discriminating real and synthetic super-resolved audio samples using embedding-based classifiersDiffusion models → Audio, speech and music
MARVEL: A Multi Agent-based Research Validator and Enabler using Large Language ModelsLLM agents and tool use → Web, search, and research agents
CodeEval: A pedagogical approach for targeted evaluation of code-trained Large Language ModelsLLM reasoning → Reasoning for agents, planning and code
The Illusion of Specialization: Unveiling the Domain-Invariant "Standing Committee" in Mixture-of-Experts ModelsMixture-of-experts → Expert routing strategies
Jailbreaking LLMs Without Gradients or Priors: Effective and Transferable AttacksLLM safety and security → Jailbreaking
GAMBIT: A Gamified Jailbreak Framework for Multimodal Large Language ModelsLLM reasoning → Reasoning for agents, planning and code +3
Sensor to Pixels: Decentralized Swarm Gathering via Image-Based Reinforcement LearningDeep reinforcement learning methods → RL for robot navigation and UAV control
Rendering Data Unlearnable by Exploiting LLM Alignment MechanismsLLM alignment / RLHF / preference optimisation → LLM preference optimization methods
Eye-Q: A Multilingual Benchmark for Visual Word Puzzle Solving and Image-to-Phrase ReasoningLLM reasoning → Multimodal reasoning +1
Metaphors are a Source of Cross-Domain Misalignment of Large Reasoning ModelsLLM reasoning → Reasoning benchmarks and evaluation
SIGMA: Scalable Spectral Insights for LLM Model CollapseSynthetic data and self-improvement for LLMs → Model collapse from synthetic data
RepoShapley: Shapley-Enhanced Context Filtering for Repository-Level Code CompletionRetrieval-augmented generation → RAG for code
Enhancing LLM Instruction Following: An Evaluation-Driven Multi-Agentic Workflow for Prompt Instructions OptimizationLLM agents and tool use → General LLM agent frameworks and benchmarks
MMErroR: A Benchmark for Erroneous Reasoning in Vision-Language ModelsLLM reasoning → Multimodal reasoning +1
LLM-Enabled Multi-Agent Systems: Empirical Evaluation and Insights into Emerging Design Patterns & ParadigmsLLM agents and tool use → General LLM agent frameworks and benchmarks
Listen to Rhythm, Choose Movements: Autoregressive Multimodal Dance Generation via Diffusion and Mamba with Decoupled Dance DatasetState-space sequence models (S4, Mamba) → Video, motion and generative Mamba models
HERO: Hierarchical Evidential Reasoning Optimization for Radiology Report Generation via Reason-then-SummarizeLLM reasoning → Clinical and biomedical applications
Ratio-Variance Regularized Policy Optimization for Efficient LLM Fine-tuningLLM reasoning → Reinforcement learning for reasoning
Why LLMs Aren't Scientists Yet: Lessons from Four Autonomous Research AttemptsLLM agents and tool use → Web, search, and research agents
VLM4VLA: Revisiting Vision-Language-Models in Vision-Language-Action ModelsMultimodal LLMs / vision-language models → Embodied and 3D vision-language models +1
Mastering the Game of Go with Self-play Experience ReplayDeep reinforcement learning methods → RL for games and combinatorial optimization
Mass Concept Erasure in Diffusion Models with Concept HierarchyDiffusion models → Robustness, privacy and adversarial aspects +1
TRYLOCK: Defense-in-Depth Against LLM Jailbreaks via Layered Preference and Representation EngineeringLLM alignment / RLHF / preference optimisation → Safety alignment and jailbreak attacks +2
STReasoner: Empowering LLMs for Spatio-Temporal Reasoning in Time Series via Spatial-Aware Reinforcement LearningLLM reasoning → Spatial and embodied reasoning
MAGMA: A Multi-Graph based Agentic Memory Architecture for AI AgentsLLM agents and tool use → Agent memory systems
LTX-2: Efficient Joint Audio-Visual Foundation ModelDiffusion models → Audio, speech and music
MalruleLib: Large-Scale Executable Misconception Reasoning with Step Traces for Modeling Student Thinking in MathematicsLLM reasoning → Mathematical reasoning
Critic-Guided Reinforcement Unlearning in Text-to-Image DiffusionDiffusion models → Diffusion and flow policies for robotics +1
UltraLogic: Enhancing LLM Reasoning through Large-Scale Data Synthesis and Bipolar Float RewardLLM reasoning → Reinforcement learning for reasoning +1
InfiAgent: An Infinite-Horizon Framework for General-Purpose Autonomous AgentsLLM agents and tool use → Agent memory systems
DIP: Dynamic In-Context Planner For Diffusion Language ModelsDiffusion models → Discrete and text diffusion
Empowering Reliable Visual-Centric Instruction Following in MLLMsMultimodal LLMs / vision-language models → MLLM training and visual instruction tuning
Software-Defined Agentic ServingLLM agents and tool use → Software engineering coding agents
Sparse Knowledge Distillation: A Mathematical Framework for Probability-Domain Temperature Scaling and Multi-Stage CompressionKnowledge distillation → General knowledge distillation methods
MemRL: Self-Evolving Agents via Runtime Reinforcement Learning on Episodic MemoryDeep reinforcement learning methods → RL training of language and agent models
AnatomiX, an Anatomy-Aware Grounded Multimodal Large Language Model for Chest X-Ray InterpretationLLM reasoning → Clinical and biomedical applications +1
Maximizing Local Entropy Where It Matters: Prefix-Aware Localized LLM UnlearningLLM safety and security → Machine unlearning
DiffBench Meets DiffAgent: End-to-End LLM-Driven Diffusion Acceleration Code GenerationDiffusion models → Diffusion foundations and sampling
WebAnchor: Anchoring Agent Planning to Stabilize Long-Horizon Web ReasoningLLM reasoning → Reasoning for agents, planning and code +1
Decoupling the Effect of Chain-of-Thought Reasoning: A Human Label Variation PerspectiveLLM reasoning → Prompting and chain-of-thought reasoning
Limited Linguistic Diversity in Embodied AI DatasetsVision-language-action models → Generalist VLA models and benchmarks
The Anatomy of Conversational Scams: A Topic-Based Red Teaming Analysis of Multi-Turn Interactions in LLMsLLM safety and security → Prompt injection and agent security
DiT-JSCC: Rethinking Deep JSCC with Diffusion Transformers and Semantic RepresentationsDiffusion models → Diffusion for compression and wireless
One Sample to Rule Them All: Extreme Data Efficiency in Multidiscipline Reasoning with Reinforcement LearningLLM reasoning → Reinforcement learning for reasoning
TGIF: Text-Guided Layer Fusion Mitigates Hallucination in Multimodal LLMsMultimodal LLMs / vision-language models → Hallucination mitigation in LVLMs
ATLAS: Verifier-Guided Adaptive Latent Activation Steering for Efficient LLM ReasoningLLM reasoning → Test-time compute and efficient thinking +1
Understanding Multi-Agent Reasoning with Large Language Models for Cartoon VQALLM reasoning → Multimodal reasoning +1
Joint Encoding of KV-Cache Blocks for Scalable LLM ServingEfficient LLM inference and serving → KV cache optimization
Do LLMs Encode Functional Importance of Reasoning Tokens?LLM reasoning → Test-time compute and efficient thinking
IBISAgent: Reinforcing Pixel-Level Visual Reasoning in MLLMs for Universal Biomedical Object Referring and SegmentationLLM reasoning → Clinical and biomedical applications
Detecting Hallucinations in Retrieval-Augmented Generation via Semantic-level Internal Reasoning GraphLLM reasoning → Hallucination detection and mitigation +1
On the Intrinsic Limits of Transformer Image Embeddings in Non-Solvable Spatial ReasoningLLM reasoning → Spatial and embodied reasoning
When the Coffee Feature Activates on Coffins: An Analysis of Feature Extraction and Steering for Mechanistic InterpretabilityMechanistic interpretability → SAE features and their applications
SOP: A Scalable Online Post-Training System for Vision-Language-Action ModelsVision-language-action models → Reinforcement learning and post-training
Flow Matching and Diffusion Models via PointNet for Generating Fluid Fields on Irregular GeometriesDiffusion models → Flow matching, time series and science
Reducing Hallucinations in LLMs via Factuality-Aware Preference LearningLLM alignment / RLHF / preference optimisation → Diffusion and multimodal model alignment
Dementia-R1: Reinforced Pretraining and Reasoning from Unstructured Clinical Notes for Real-World Dementia PrognosisLLM reasoning → Clinical and biomedical applications
Adaptive Control of Unknown Linear Switched Systems via Policy Gradient MethodsDeep reinforcement learning methods → RL for games and combinatorial optimization
In-Context Reinforcement Learning through Bayesian Fusion of Context and Value PriorDeep reinforcement learning methods → Deep RL algorithms and representations
SentGraph: Hierarchical Sentence Graph for Multi-hop Retrieval-Augmented Question AnsweringRetrieval-augmented generation → Knowledge-graph and structured-knowledge RAG
JPU: Bridging Jailbreak Defense and Unlearning via On-Policy Path RectificationLLM alignment / RLHF / preference optimisation → Safety alignment and jailbreak attacks +1
Large Reasoning Models Are (Not Yet) Multilingual Latent ReasonersLLM reasoning → Reasoning benchmarks and evaluation
Stable-RAG: Mitigating Retrieval-Permutation-Induced Hallucinations in Retrieval-Augmented GenerationRetrieval-augmented generation → Hallucination detection in RAG
Towards Faithful Reasoning in Comics for Small MLLMsLLM reasoning → Multimodal reasoning +1
Mechanistic Interpretability of Large-Scale Counting in LLMs through a System-2 StrategyMechanistic interpretability → Steering and interpreting LLM reasoning
LAMS-Edit: Latent and Attention Mixing with Schedulers for Improved Content Preservation in Diffusion-Based Image and Style EditingDiffusion models → Image generation, text-to-image and editing
P-Check: Advancing Personalized Reward Model via Learning to Generate Dynamic ChecklistLLM alignment / RLHF / preference optimisation → Human values and AI feedback alignment
Interpretable All-Type Audio Deepfake Detection with Audio LLMs via Frequency-Time Reinforcement LearningLLM reasoning → Audio-language model reasoning
Mechanistic Knobs in LLMs: Retrieving and Steering High-Order Semantic Features via Sparse AutoencodersMechanistic interpretability → SAE features and their applications
Correct, Concise and Complete: Multi-stage Training For Adaptive ReasoningLLM reasoning → Test-time compute and efficient thinking
Reliability-Aware Adaptive Self-Consistency for Efficient Sampling in LLM ReasoningLLM reasoning → Test-time compute and efficient thinking
Rationale-Grounded In-Context Learning for Time Series Reasoning with Multimodal Large Language ModelsLLM reasoning → Time series reasoning +1
MoE Adapter for Large Audio Language Models: Sparsity, Disentanglement, and Gradient-Conflict-FreeMixture-of-experts → Multimodal MoE models
Enhancing Multilingual RAG Systems with Debiased Language Preference-Guided Query FusionRetrieval-augmented generation → Generation and context integration +1
Batch-of-Thought: Cross-Instance Learning for Enhanced LLM ReasoningLLM reasoning → Reasoning for agents, planning and code
XLSR-MamBo: Scaling the Hybrid Mamba-Attention Backbone for Audio Deepfake DetectionState-space sequence models (S4, Mamba) → Speech and audio SSMs
MixTTE: Multi-Level Mixture-of-Experts for Scalable and Adaptive Travel Time EstimationMixture-of-experts → MoE for time series and graphs
SastBench: A Benchmark for Testing Agentic SAST TriageLLM agents and tool use → Agent threats, guardrails and governance
Zoom-IQA: Image Quality Assessment with Reliable Region-Aware ReasoningLLM reasoning → Multimodal reasoning
RAL2M: Retrieval Augmented Learning-To-Match Against Hallucination in Compliance-Guaranteed Service SystemsRetrieval-augmented generation → Hallucination detection in RAG
Logical Phase Transitions: Understanding Collapse in LLM Logical ReasoningLLM reasoning → Reasoning benchmarks and evaluation
Bridging Mechanistic Interpretability and Prompt Engineering with Gradient Ascent for Interpretable Persona ControlMechanistic interpretability → Activation steering methods

All 357