Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer

2019/10/23 by Colin Raffel, Noam Shazeer, Raffel, Colin +16 · 4 voices · 1173 citations
Computer Science · #Multimodal Machine Learning Applications #Natural Language Processing Techniques #Topic Modeling #cs.CL #cs.LG #stat.ML

paper · pdf · doi:10.48550/arxiv.1910.10683

openalex publication_date 2019/10/23 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

Transfer learning, where a model is first pre-trained on a data-rich task before being fine-tuned on a downstream task, has emerged as a powerful technique in natural language processing (NLP). The effectiveness of transfer learning has given rise to a diversity of approaches, methodology, and practice. In this paper, we explore the landscape of transfer learning techniques for NLP by introducing a unified framework that converts all text-based language problems into a text-to-text format. Our systematic study compares pre-training objectives, architectures, unlabeled data sets, transfer approaches, and other factors on dozens of language understanding tasks. By combining the insights from our exploration with scale and our new ``Colossal Clean Crawled Corpus'', we achieve state-of-the-art results on many benchmarks covering summarization, question answering, text classification, and more. To facilitate future work on transfer learning for NLP, we release our data set, pre-trained models, and code.

Cited by

• Merge before Forget: A Single LoRA Continual Learning via Continual Merging
• Identifying social bots via heterogeneous motifs based on Naïve Bayes model
• RealCamo: Boosting Real Camouflage Synthesis with Layout Controls and Textual-Visual Guidance
• ShapeR: Robust Conditional 3D Shape Generation from Casual Captures
• Bridging Global Intent with Local Details: A Hierarchical Representation Approach for Semantic Validation in Text-to-SQL
• Visual Autoregressive Modelling for Monocular Depth Estimation
• SPECTRE: Spectral Pre-training Embeddings with Cylindrical Temporal Rotary Position Encoding for Fine-Grained sEMG-Based Movement Decoding
• VideoScaffold: Elastic-Scale Visual Hierarchies for Streaming Video Understanding in MLLMs
• Towards Signboard-Oriented Visual Question Answering: ViSignVQA Dataset, Method and Benchmark
• RayRoPE: Projective Ray Positional Encoding for Multi-view Attention
• VIPER: Visual In-Context Physics Reasoning for Physically Plausible Video Generation
• Selecting Language Models for Social Science: Start Small, Start Open, and Validate
• Towards a Relevance Posterior in Neural Information Access
• BERT-based Models vs. Large Language Models for Low-Resource Named Entity Recognition: A Comparative Study on Marathi
• The JEPA Paradox in Language: The Geometry of Linguistic Alternatives
• Neonatal Hypoxic-ischaemic Encephalopathy Classification from the EEG and HRV Signals Using a Conformer based Masked Autoencoder
• Accelerating Language Model Workflows with Prompt Choreography
• Enhancing Code Understanding for Impact Analysis by Combining Transformers and Program Dependence Graphs
• Latent-LoRA: Compact Latent-Space Adapters with Gradient-Free Routing for Continual Learning
• Where meaning lives: Layer-wise accessibility of psycholinguistic features in encoder and decoder language models
• PriSAR: 3D Geometric-Prior-Guided Diffusion for Parameter-Controlled SAR Image Generation
• Foundation Models and Fine-Tuning: Toward a New Generation of Models for Time Series Forecasting
• Gnowsis: Multimodal multitask learning for oral proficiency assessments
• AutoFed: Personalized Federated Traffic Prediction via Adaptive Prompt
• VaLiDRec: Variable-Length LLM-Aligned Semantic IDs for Generative Recommendation
• Architectural Backdoors in Vision-Language Model Supply Chains via Representation Steering
• Hierarchical Grading in Large Language Models
• Open Your Model’s Eyes: Video and Context-Aware Multimodal Backchannel Prediction
• Similarity All The Way Up: Multilingual Generalization in LLMs Relies on Language-Level Similarity Structures
• Reading Without a Reader: Large Language Models Collapse Reading and Writing into a Single Entangled Code
• T2LDM++: A Self-Conditioned Representation Guided Diffusion Model for Realistic Text-to-LiDAR Scene Generation
• FARM: Find Anything using Relational Spatial Memory
• Neural Machine Translation for Low-Resource Tangkhul--English
• GrocLM: Grocery Category Recommendation in E-Commerce with Large Language Models
• Evolving from Lessons: Skill-Augmented Table Graph Reasoning for Operation-wise Table Question Answering
• Schema-Aware Localisation (SAL): Live Schema Grounding and Hallucination Validation for Oracle NL2SQL
• Same Question, Different Answers: Evaluating LLM Reliability Beyond Accuracy
• FlashEvaluator: Expanding Search Space with Parallel Sequence-Level Evaluation
• Adapting a Text-to-Audio Model for Room Impulse Response Generation
• DPAR: Dynamic Patchification for Efficient Autoregressive Visual Generation
• Approximation Capabilities of Feedforward Neural Networks with GELU Activations
• Cross-Semantic Transfer Learning for High-Dimensional Linear Regression
• Rethinking Output Alignment For 1-bit Post-Training Quantization of Large Language Models
• CEMG: Collaborative-Enhanced Multimodal Generative Recommendation
• RefineBridge: Generative Bridge Models Improve Financial Forecasting by Foundation Models
• Enabling Conversational Behavior Reasoning Capabilities in Full-Duplex Speech
• ACD: Direct Conditional Control for Video Diffusion Models via Attention Supervision
• Beyond Pixel Simulation: Pathology Image Generation via Diagnostic Semantic Tokens and Prototype Control
• When LLMs fall short in Deductive Coding: Model Comparison and Human AI Collaboration Workflow Design
• Artificial or Just Artful? Do LLMs Bend the Rules in Programming?
• TokSuite: Measuring the Impact of Tokenizer Choice on Language Model Behavior
• On the Effectiveness of Instruction-Tuning Local LLMs for Identifying Software Vulnerabilities
• Reason2Decide: Rationale-Driven Multi-Task Learning
• QuarkAudio Technical Report
• Making Large Language Models Efficient Dense Retrievers
• Diacritic Restoration for Low-Resource Indigenous Languages: Case Study with Bribri and Cook Islands Māori
• Event Extraction in Large Language Model
• OmniMoGen: Unifying Human Motion Generation via Learning from Interleaved Text-Motion Instructions
• AraMix: Recycling, Refiltering, and Deduplicating to Deliver the Largest Arabic Pretraining Corpus
• A Comparative Study of Light-weight Language Models for PII Masking and their Deployment for Real Conversational Texts
• A Multi-agent Text2SQL Framework using Small Language Models and Execution Feedback
• A Data-Centric Approach to Generalizable Speech Deepfake Detection
• AnyTask: an Automated Task and Data Generation Framework for Advancing Sim-to-Real Policy Learning
• ShareChat: A Dataset of Chatbot Conversations in the Wild
• Bridging Natural Language and Formal Specification--Automated Translation of Software Requirements to LTL via Hierarchical Semantics Decomposition Using LLMs
• Easy Adaptation: An Efficient Task-Specific Knowledge Injection Method for Large Models in Resource-Constrained Environments
• Exploiting ID-Text Complementarity via Ensembling for Sequential Recommendation
• Toward Ethical AI Through Bayesian Uncertainty in Neural Question Answering
• Smoothing DiLoCo with Primal Averaging for Faster Training of LLMs
• Bandwidth-Efficient Adaptive Mixture-of-Experts via Low-Rank Compensation
• VIVA: VLM-Guided Instruction-Based Video Editing with Reward Optimization
• Grammar-Forced Translation of Natural Language to Temporal Logic using LLMs
• GinSign: Grounding Natural Language Into System Signatures for Temporal Logic Translation
• Yuan-TecSwin: A text conditioned Diffusion model with Swin-transformer blocks
• Factorized Video Generation: Decoupling Scene Construction and Temporal Synthesis in Text-to-Video Diffusion Models
• The Evolution of Reranking Models in Information Retrieval: From Heuristic Methods to Large Language Models
• DualGuard: Dual-stream Large Language Model Watermarking Defense against Paraphrase and Spoofing Attack
• PAACE: A Plan-Aware Automated Agent Context Engineering Framework
• mimic-video: Video-Action Models for Generalizable Robot Control Beyond VLAs
• GRAN-TED: Generating Robust, Aligned, and Nuanced Text Embedding for Diffusion Models
• MiVLA: Towards Generalizable Vision-Language-Action Model with Human-Robot Mutual Imitation Pre-training
• Beyond Fast and Slow: Cognitive-Inspired Elastic Reasoning for Large Language Models
• Prompt Repetition Improves Non-Reasoning LLMs
• Spatia: Video Generation with Updatable Spatial Memory
• T5Gemma 2: Seeing, Reading, and Understanding Longer
• Per-Axis Weight Deltas for Frequent Model Updates
• Dual-objective Language Models: Training Efficiency Without Overfitting
• Effect of Document Packing on the Latent Multi-Hop Reasoning Capabilities of Large Language Models
• OmniGen: Unified Multimodal Sensor Generation for Autonomous Driving
• DRAW2ACT: Turning Depth-Encoded Trajectories into Robotic Demonstration Videos
• Explainable Ethical Assessment on Human Behaviors by Generating Conflicting Social Norms
• DTop-p MoE: Sparsity-Controlled Dynamic Top-p MoE for Foundation Model Pre-training
• MoLingo: Motion-Language Alignment for Text-to-Human Motion Generation
• Advancing Bangla Machine Translation Through Informal Datasets
• LINA: Learning INterventions Adaptively for Physical Alignment and Generalization in Diffusion Models
• MMDrive: Interactive Scene Understanding Beyond Vision with Multi-representational Fusion
• Alada: Alternating Adaptation of Momentum Method for Memory-Efficient Matrix Optimization
• SkipCat: Rank-Maximized Low-Rank Compression of Large Language Models via Shared Projection and Block Skipping
• Informing Acquisition Functions via Foundation Models for Molecular Discovery
• SPAR: Session-based Pipeline for Adaptive Retrieval on Legacy File Systems
• PrahokBART: A Pre-trained Sequence-to-Sequence Model for Khmer Natural Language Generation
• Robust Motion Generation using Part-level Reliable Data from Videos
• Resting Neurons, Active Insights: Improving Input Sparsification for Large Language Models
• FuXi-γ: Efficient Sequential Recommendation with Exponential-Power Temporal Encoder and Diagonal-Sparse Positional Mechanism
• EditMGT: Unleashing Potentials of Masked Generative Transformers in Image Editing
• AdaGradSelect: An adaptive gradient-guided layer selection method for efficient fine-tuning of SLMs
• PhraseVAE and PhraseLDM: Latent Diffusion for Full-Song Multitrack Symbolic Music Generation
• AdaSD: Adaptive Speculative Decoding for Efficient Language Model Inference
• Multi-Intent Spoken Language Understanding: Methods, Trends, and Challenges
• REST: Diffusion-based Real-time End-to-end Streaming Talking Head Generation via ID-Context Caching and Asynchronous Streaming Distillation
• SciLaD: A Large-Scale, Transparent, Reproducible Dataset for Natural Scientific Language Processing
• An Anatomy of Vision-Language-Action Models: From Modules to Milestones and Challenges
• TAO-Net: Two-stage Adaptive OOD Classification Network for Fine-grained Encrypted Traffic Classification
• Stronger Normalization-Free Transformers
• SparseSwaps: Tractable LLM Pruning Mask Refinement at Scale
• Textual Data Bias Detection and Mitigation -- An Extensible Pipeline with Experimental Evaluation
• Learning by Analogy: A Causal Framework for Composition Generalization
• Zero-shot Adaptation of Stable Diffusion via Plug-in Hierarchical Degradation Representation for Real-World Super-Resolution
• Watermarks for Language Models via Probabilistic Automata
• Beyond the Black Box: Identifiable Interpretation and Control in Generative Models via Causal Minimality
• Defining Cost Function of Steganography with Large Language Models
• Interpreto: An Explainability Library for Transformers
• Learning Patient-Specific Disease Dynamics with Latent Flow Matching for Longitudinal Imaging Generation
• Food Image Generation on Multi-Noun Categories
• UniLayDiff: A Unified Diffusion Transformer for Content-Aware Layout Generation
• Revisiting the Scaling Properties of Downstream Metrics in Large Language Model Training
• Financial News Summarization: Can extractive methods still offer a true alternative to LLMs?
• The Erosion of LLM Signatures: Can We Still Distinguish Human and LLM-Generated Scientific Ideas After Iterative Paraphrasing?
• Beyond the Noise: Aligning Prompts with Latent Representations in Diffusion Models
• Attention is All You Need to Defend Against Indirect Prompt Injection Attacks in LLMs
• Luxical: High-Speed Lexical-Dense Text Embeddings
• ValuePilot: A Two-Phase Framework for Value-Driven Decision-Making
• Guiding What Not to Generate: Automated Negative Prompting for Text-Image Alignment
• LIME: Making LLM Data More Efficient with Linguistic Metadata Embeddings
• SETUP: Sentence-level English-To-Uniform Meaning Representation Parser
• FOAM: Blocked State Folding for Memory-Efficient LLM Training
• A Patient-Doctor-NLP-System to contest inequality for less privileged
• On Memory: A comparison of memory mechanisms in world models
• VideoVLA: Video Generators Can Be Generalizable Robot Manipulators
• Teaching Language Models Mechanistic Explainability Through Arrow-Pushing
• EgoEdit: Dataset, Real-Time Streaming Model, and Benchmark for Egocentric Video Editing
• KQ-SVD: Compressing the KV Cache with Provable Guarantees on Attention Fidelity
• From Text to Returns: Using Large Language Models for Mutual Fund Portfolio Optimization and Risk-Adjusted Allocation
• Scaling and Transferability of Annealing Strategies in Large Language Model Training
• 2K-Characters-10K-Stories: A Quality-Gated Stylized Narrative Dataset with Disentangled Control and Sequence Consistency
• MIND: Multi-rationale INtegrated Discriminative Reasoning Framework for Multi-modal Large Models
• Light-X: Generative 4D Video Rendering with Camera and Illumination Control
• ShaRP: SHAllow-LayeR Pruning for Video Large Language Models Acceleration
• Aligned but Stereotypical? The Hidden Influence of System Prompts on Social Bias in LVLM-Based Text-to-Image Models
• GeoPE:A Unified Geometric Positional Embedding for Structured Tensors
• STELLA: Guiding Large Language Models for Time Series Forecasting with Semantic Abstractions
• OsmT: Bridging OpenStreetMap Queries and Natural Language with Open-source Tag-aware Language Models
• Context-Aware Mixture-of-Experts Inference on CXL-Enabled GPU-NDP Systems
• RapidUn: Influence-Driven Parameter Reweighting for Efficient Large Language Model Unlearning
• Stable Signer: Hierarchical Sign Language Generative Model
• Enhancing Instruction-Following Capabilities in Seq2Seq Models: DoLA Adaptations for T5
• Tutorial on Large Language Model-Enhanced Reinforcement Learning for Wireless Networks
• AlignCheck: a Semantic Open-Domain Metric for Factual Consistency Assessment
• RoboScape-R: Unified Reward-Observation World Models for Generalizable Robotics Training via RL
• CookAnything: A Framework for Flexible and Consistent Multi-Step Recipe Image Generation
• Data-Free Pruning of Self-Attention Layers in LLMs
• FloodDiffusion: Tailored Diffusion Forcing for Streaming Motion Generation
• MarkTune: Improving the Quality-Detectability Trade-off in Open-Weight LLM Watermarking
• Video2Act: A Dual-System Video Diffusion Policy with Robotic Spatio-Motional Modeling
• MultiShotMaster: A Controllable Multi-Shot Video Generation Framework
• Fairy2i: Training Complex LLMs from Real LLMs with All Parameters in \± 1, ± i\
• Network Self-Configuration based on Fine-Tuned Small Language Models
• PEFT-Factory: Unified Parameter-Efficient Fine-Tuning of Autoregressive Large Language Models
• ESACT: An End-to-End Sparse Accelerator for Compute-Intensive Transformers via Local Similarity
• Offloading Artificial Intelligence Workloads across the Computing Continuum by means of Active Storage Systems
• Parameter-Efficient Subspace Optimization for LLM Fine-Tuning
• Think Before You Prune: Self-Reflective Structured Pruning for Reasoning Language Models
• Low-Rank Prehab: Preparing Neural Networks for SVD Compression
• Reconstructing Multi-Scale Physical Fields from Extremely Sparse Measurements with an Autoencoder-Diffusion Cascade
• LAURA: Enhancing Code Review Generation with Context-Enriched Retrieval-Augmented LLM
• TokenPure: Watermark Removal through Tokenized Appearance and Structural Guidance
• DCText: Scheduled Attention Masking for Visual Text Generation via Divide-and-Conquer Strategy
• On The Finetuning of MLIPs Through the Lens of Iterated Maps With BPTT
• Table as a Modality for Large Language Models
• HBLLM: Wavelet-Enhanced High-Fidelity 1-Bit Quantization for LLMs
• WaterSearch: A Quality-Aware Search-based Watermarking Framework for Large Language Models
• Financial Text Classification Based On rLoRA Finetuning On Qwen3-8B model
• Low-Bitrate Video Compression through Semantic-Conditioned Diffusion
• Simplex-Optimized Hybrid Ensemble for Large Language Model Text Detection Under Generative Distribution Drif
• Comparative Analysis of 47 Context-Based Question Answer Models Across 8 Diverse Datasets
• Video-R2: Reinforcing Consistent and Grounded Reasoning in Multimodal Language Models
• AnyTalker: Scaling Multi-Person Talking Video Generation with Interactivity Refinement
• Resolving Conflicts in Lifelong Learning via Aligning Updates in Subspaces
• Tourism Question Answer System in Indian Language using Domain-Adapted Foundation Models
• Decoding the Past: Explainable Machine Learning Models for Dating Historical Texts
• Language-conditioned world model improves policy generalization by reading environmental descriptions
• Towards Improving Interpretability of Language Model Generation through a Structured Knowledge Discovery Approach
• Dripper: Token-Efficient Main HTML Extraction with a Lightweight LM
• CoFiRec: Coarse-to-Fine Tokenization for Generative Recommendation
• Exploring Performance Variations in Finetuned Translators of Ultra-Low Resource Languages: Do Linguistic Differences Matter?
• TraceGen: World Modeling in 3D Trace Space Enables Learning from Cross-Embodiment Videos
• SingleQuant: Efficient Quantization of Large Language Models in a Single Pass
• Beyond URLs: Metadata Diversity and Position for Efficient LLM Pretraining
• Odin: Oriented Dual-module Integration for Text-rich Network Representation Learning
• Uni-Hema: Unified Model for Digital Hematopathology
• A Systematic Study of In-the-Wild Model Merging for Large Language Models
• Infinite-Story: A Training-Free Consistent Text-to-Image Generation
• 3MDiT: Unified Tri-Modal Diffusion Transformer for Text-Driven Synchronized Audio-Video Generation
• PEFT-Bench: A Parameter-Efficient Fine-Tuning Methods Benchmark
• CameraMaster: Unified Camera Semantic-Parameter Control for Photography Retouching
• Towards Audio Token Compression in Large Audio Language Models
• CafeQ: Calibration-free Quantization via Learned Transformations and Adaptive Rounding
• Chatty-KG: A Multi-Agent AI System for On-Demand Conversational Question Answering over Knowledge Graphs
• On the Origin of Algorithmic Progress in AI
• Text-Guided Semantic Image Encoder
• Dynamical Properties of Tokens in Self-Attention and Effects of Positional Encoding
• Copyright Detection in Large Language Models: An Ethical Approach to Generative AI Development
• CrossEarth-Gate: Fisher-Guided Adaptive Tuning Engine for Efficient Adaptation of Cross-Domain Remote Sensing Semantic Segmentation
• HiCoGen: Hierarchical Compositional Text-to-Image Generation in Diffusion Models via Reinforcement Learning
• GigaWorld-0: World Models as Data Engine to Empower Embodied AI
• Cisco Time Series Model Technical Report
• Decoupling and Damping: Structurally-Regularized Gradient Matching for Multimodal Graph Condensation
• Cross Domain Evaluation of Multimodal Chain-of-Thought Reasoning of different datasets into the Amazon CoT Framework
• ABM-LoRA: Activation Boundary Matching for Fast Convergence in Low-Rank Adaptation
• FastForward Pruning: Efficient LLM Pruning via Single-Step Reinforcement Learning
• FineXtrol: Controllable Motion Generation via Fine-Grained Text
• Think Before You Prune: Selective Self-Generated Calibration for Pruning Large Reasoning Models
• Large Language Models for the Summarization of Czech Documents: From History to the Present
• Large-Scale In-Game Outcome Forecasting for Match, Team and Players in Football using an Axial Transformer Neural Network
• Now You See It, Now You Don't - Instant Concept Erasure for Safe Text-to-Image and Video Generation
• ProxT2I: Efficient Reward-Guided Text-to-Image Generation via Proximal Diffusion
• A symbolic Perl algorithm for the unification of Nahuatl word spellings
• IRSDA: An Agent-Orchestrated Framework for Enterprise Intrusion Response
• DELTA: Language Diffusion-based EEG-to-Text Architecture
• An Invariant Latent Space Perspective on Language Model Inversion
• Zero-Shot Video Deraining with Video Diffusion Models
• Foundations of Artificial Intelligence Frameworks: Notion and Limits of AGI
• A Systematic Study of Compression Ordering for Large Language Models
• CrossJEPA: Cross-Modal Joint-Embedding Predictive Architecture for Efficient 3D Representation Learning from 2D Images
• SmolKalam: Ensemble Quality-Filtered Translation at Scale for High Quality Arabic Post-Training Data
• Generative Adversarial Post-Training Mitigates Reward Hacking in Live Human-AI Music Interaction
• Plan-X: Instruct Video Generation via Semantic Planning
• Blu-WERP (Web Extraction and Refinement Pipeline): A Scalable Pipeline for Preprocessing Large Language Model Datasets
• Layer-Wise High-Impact Parameter Ratio Optimization in Post-Training Quantization for Large Language Models
• Video-R4: Reinforcing Text-Rich Video Reasoning with Visual Rumination
• Masked-and-Reordered Self-Supervision for Reinforcement Learning from Verifiable Rewards
• Intervene-All-Paths: Unified Mitigation of LVLM Hallucinations across Alignment Formats
• R2Q: Towards Robust 2-Bit Large Language Models via Residual Refinement Quantization
• Learning to Compress: Unlocking the Potential of Large Language Models for Text Representation
• Don't Learn, Ground: A Case for Natural Language Inference with Visual Grounding
• CREST: Improving Interpretability and Effectiveness of Troubleshooting at Ericsson through Criterion-Specific Trouble Report Retrieval
• Energy Scaling Laws for Diffusion Models: Quantifying Compute and Carbon Emissions in Image Generation
• Supervised Fine Tuning of Large Language Models for Domain Specific Knowledge Graph Construction:A Case Study on Hunan's Historical Celebrities
• Developmental Atlas of Attention Head Specialization: Spacing, Stranding, and the Capacity Tax of BPE Tokenization
• Exploring Scientific Debt: Harnessing AI for SATD Identification in Scientific Software
• Adaptive Layer-Wise Transformations for Post-Training Quantization of Large Language Models
• When Structure Doesn't Help: LLMs Do Not Read Text-Attributed Graphs as Effectively as We Expected
• You Only Forward Once: An Efficient Compositional Judging Paradigm
• LAOF: Robust Latent Action Learning with Optical Flow Constraints
• Walrus: A Cross-Domain Foundation Model for Continuum Dynamics
• NLP Datasets for Idiom and Figurative Language Tasks
• AskDB: An LLM Agent for Natural Language Interaction with Relational Databases
• UniFit: Towards Universal Virtual Try-on with MLLM-Guided Semantic Alignment
• Tokenize Once, Recommend Anywhere: Unified Item Tokenization for Multi-domain LLM-based Recommendation
• Entropy-Based Measurement of Value Drift and Alignment Work in Large Language Models
• What Does It Take to Be a Good AI Research Agent? Studying the Role of Ideation Diversity
• IPR-1: Interactive Physical Reasoner
• Text2Loc++: Generalizing 3D Point Cloud Localization from Natural Language
• SplitFlux: Learning to Decouple Content and Style from a Single Image
• PocketLLM: Ultimate Compression of Large Language Models via Meta Networks
• Insert In Style: A Zero-Shot Generative Framework for Harmonious Cross-Domain Object Composition
• Effective Code Membership Inference for Code Completion Models via Adversarial Prompts
• UniHOI: Unified Human-Object Interaction Understanding via Unified Token Space
• Bias in, Bias out: Annotation Bias in Multilingual Large Language Models
• Gradient-descent methods for quantum detector tomography
• Masked IRL: LLM-Guided Reward Disambiguation from Demonstrations and Language
• Let Language Constrain Geometry: Vision-Language Models as Semantic and Spatial Critics for 3D Generation
• ArbESC+: Arabic Enhanced Edit Selection System Combination for Grammatical Error Correction Resolving conflict and improving system combination in Arabic GEC
• Semantic Context Matters: Improving Conditioning for Autoregressive Models
• DEVAL: A Framework for Evaluating and Improving the Derivation Capability of Large Language Models
• Foundational Question Generation for Video Question Answering via an Embedding-Integrated Approach
• Intermediate N-Gramming: Deterministic and Fast N-Grams For Large N and Large Datasets
• CreBench: Human-Aligned Creativity Evaluation from Idea to Process to Product
• Scalable and Efficient Large-Scale Log Analysis with LLMs: An IT Software Support Case Study
• CorrectAD: A Self-Correcting Agentic System to Improve End-to-end Planning in Autonomous Driving
• Translation Entropy: A Statistical Framework for Evaluating Translation Systems
• NeuroLex: A Lightweight Domain Language Model for EEG Report Understanding and Generation
• ExplicitLM: Decoupling Knowledge from Parameters via Explicit Memory Banks
• Multimodal Large Language Models as Image Classifiers
• CAT-ID²: Category-Tree Integrated Document Identifier Learning for Generative Retrieval In E-commerce
• Multivariate Diffusion Transformer with Decoupled Attention for High-Fidelity Mask-Text Collaborative Facial Generation
• Designed to Spread: Generative Approaches to Enhance Information Diffusion
• HiGFA: Hierarchical Guidance for Fine-grained Data Augmentation with Diffusion Models
• Do LLMs and Humans Find the Same Questions Difficult? A Case Study on Japanese Quiz Answering
• Mixture of States: Routing Token-Level Dynamics for Multimodal Generation
• MAVIS: A Benchmark for Multimodal Source Attribution in Long-form Visual Question Answering
• OAD-Promoter: Enhancing Zero-shot VQA using Large Language Models with Object Attribute Description
• GeoMVD: Geometry-Enhanced Multi-View Generation Model Based on Geometric Information Extraction
• KVSwap: Disk-aware KV Cache Offloading for Long-Context On-device Inference
• Context-Emotion Aware Therapeutic Dialogue Generation: A Multi-component Reinforcement Learning Approach to Language Models for Mental Health Support
• BhashaKritika: Building Synthetic Pretraining Data at Scale for Indic Languages
• Large Language Models and 3D Vision for Intelligent Robotic Perception and Autonomy
• Hindsight Distillation Reasoning with Knowledge Encouragement Preference for Knowledge-based Visual Question Answering
• Stroke Modeling Enables Vectorized Character Generation with Large Vectorized Glyph Model
• Phantom Menace: Exploring and Enhancing the Robustness of VLA Models Against Physical Sensor Attacks
• Improving LLM's Attachment to External Knowledge In Dialogue Generation Tasks Through Entity Anonymization
• Unveiling the Impact of Data and Model Scaling on High-Level Control for Humanoid Robots
• Text2SQL-Flow: A Robust SQL-Aware Data Augmentation Framework for Text-to-SQL
• ChEmREF: Evaluating Language Model Readiness for Chemical Emergency Response
• TermGPT: Multi-Level Contrastive Fine-Tuning for Terminology Adaptation in Legal and Financial Domain
• Making Every Head Count: Sparse Attention Without the Speed-Performance Trade-off
• Selective Sinkhorn Routing for Improved Sparse Mixture of Experts
• CMI-MTL: Cross-Mamba interaction based multi-task learning for medical visual question answering
• Not Everything That Counts Can Be Counted: A Case for Safe Qualitative AI
• CleverBirds: A Multiple-Choice Benchmark for Fine-grained Human Knowledge Tracing
• Introducing A Bangla Sentence - Gloss Pair Dataset for Bangla Sign Language Translation and Research
• Why Larger Models Learn More: Effects of Capacity, Interference, and Rare-Task Retention
• TurkEmbed: Turkish Embedding Model on NLI & STS Tasks
• VideoChain: A Transformer-Based Framework for Multi-hop Video Question Generation
• JobSphere: An AI-Powered Multilingual Career Copilot for Government Employment Platforms
• Determinism of Randomness: Prompt-Residual Seed Shaping for Diffusion Generation
• A Unified Geometric Field Theory Framework for Transformers: From Manifold Embeddings to Kernel Modulation
• Multi-Granularity Mutual Refinement Network for Zero-Shot Learning
• Laytrol: Preserving Pretrained Knowledge in Layout Control for Multimodal Diffusion Transformers
• WaterMod: Modular Token-Rank Partitioning for Probability-Balanced LLM Watermarking
• Towards General Auditory Intelligence: Large Multimodal Models for Machine Listening and Speaking
• Majority Rules: LLM Ensemble is a Winning Approach for Content Categorization
• A Circular Argument : Does RoPE need to be Equivariant for Vision?
• oboro: Text-to-Image Synthesis on Limited Data using Flow-based Diffusion Transformer with MMH Attention
• TurkEmbed4Retrieval: Turkish Embedding Model for Retrieval Task
• AraFinNews: Arabic Financial Summarisation with Domain-Adapted LLMs
• Beyond English: Toward Inclusive and Scalable Multilingual Machine Translation with LLMs
• Generating an Image From 1,000 Words: Enhancing Text-to-Image With Structured Captions
• How Do Data Owners Say No? A Case Study of Data Consent Mechanisms in Web-Scraped Vision-Language AI Training Datasets
• Rethinking Parameter Sharing as Graph Coloring for Structured Compression
• FedRW: Efficient Privacy-Preserving Data Reweighting for Enhancing Federated Learning of Language Models
• LLM Optimization Unlocks Real-Time Pairwise Reranking
• A Decentralized Retrieval Augmented Generation System with Source Reliabilities Secured on Blockchain
• Route Experts by Sequence, not by Token
• Vocabulary In-Context Learning in Transformers: Benefits of Positional Encoding
• Reaction Prediction via Interaction Modeling of Symmetric Difference Shingle Sets
• Seq2Seq Models Reconstruct Visual Jigsaw Puzzles without Seeing Them
• TinyChemVL: Advancing Chemical Vision-Language Models via Efficient Visual Token Reduction and Complex Reaction Tasks
• Alignment-Constrained Dynamic Pruning for LLMs: Identifying and Preserving Alignment-Critical Circuits
• FLEX: Continuous Agent Evolution via Forward Learning from Experience
• LLaDA-Rec: Discrete Diffusion for Parallel Semantic ID Generation in Generative Recommendation
• L2T-Hyena: Enhancing State-Space Models with an Adaptive Learn-to-Teach Framework
• MOSS: Efficient and Accurate FP8 LLM Training with Microscaling and Automatic Scaling
• TabDistill: Distilling Transformers into Neural Nets for Few-Shot Tabular Classification
• A Representation Sharpening Framework for Zero Shot Dense Retrieval
• ManufactuBERT: Efficient Continual Pretraining for Manufacturing
• Reasoning-Guided Claim Normalization for Noisy Multilingual Social Media Posts
• Search Is Not Retrieval: Decoupling Semantic Matching from Contextual Assembly in RAG
• REFLEX: Reference-Free Evaluation of Log Summarization via Large Language Model Judgment
• PuzzleMoE: Efficient Compression of Large Mixture-of-Experts Models via Sparse Expert Merging and Bit-packed inference
• PromptSep: Generative Audio Separation via Multimodal Prompting
• RISE-T2V: Rephrasing and Injecting Semantics with LLM for Expansive Text-to-Video Generation
• Reusing Pre-Training Data at Test Time is a Compute Multiplier
• ScaleDL: Towards Scalable and Efficient Runtime Prediction for Distributed Deep Learning Workloads
• E-CARE: An Efficient LLM-based Commonsense-Augmented Framework for E-Commerce
• DartQuant: Efficient Rotational Distribution Calibration for LLM Quantization
• MIDI-LLM: Adapting Large Language Models for Text-to-MIDI Music Generation
• Promoting Sustainable Web Agents: Benchmarking and Estimating Energy Consumption through Empirical and Theoretical Analysis
• OMPILOT: Harnessing Transformer Models for Auto Parallelization to Shared Memory Computing Paradigms
• Diffusion Language Models are Super Data Learners
• Divide, Cache, Conquer: Dichotomic Prompting for Efficient Multi-Label LLM-Based Classification
• Data-Efficient Adaptation and a Novel Evaluation Method for Aspect-based Sentiment Analysis
• Targeted Error Correction in Knowledge Distillation: Small Language Models Surpass GPT
• Dynamic Reflections: Probing Video Representations with Text Alignment
• AyurParam: A State-of-the-Art Bilingual Language Model for Ayurveda
• ReleaseEval: A Benchmark for Evaluating Language Models in Automated Release Note Generation
• Memory-Efficient Training with In-Place FFT Implementation
• Random Initialization of Gated Sparse Adapters
• HPLT 3.0: Very Large-Scale Multilingual Resources for LLMs and MT. Mono- and Bi-lingual Data, Multilingual Evaluation, and Pre-Trained Models
• Reviving Stale Updates: Data-Free Knowledge Distillation for Asynchronous Federated Learning
• Air Pollution Forecasting in Bucharest
• Listwise Preference Diffusion Optimization for User Behavior Trajectories Prediction
• ID-Crafter: VLM-Grounded Online RL for Compositional Multi-Subject Video Generation
• Diff4Splat: Controllable 4D Scene Generation with Latent Dynamic Reconstruction Models
• Isotropic Curvature Model for Understanding Deep Learning Optimization: Is Gradient Orthogonalization Optimal?
• BiSparse-AAS: Bilinear Sparse Attention and Adaptive Spans Framework for Scalable and Efficient Text Summarization
• Generative Semantic Coding for Ultra-Low Bitrate Visual Communication and Analysis
• Probability-Biased Attention over Directed Bipartite Graphs for Long-Tail ICD Coding
• E-MMDiT: Revisiting Multimodal Diffusion Transformer Design for Fast Image Synthesis under Limited Resources
• EBT-Policy: Energy Unlocks Emergent Physical Reasoning Capabilities
• Reasoning Up the Instruction Ladder for Controllable Language Models
• Mixture-of-Transformers Learn Faster: A Theoretical Study on Classification Problems
• Integrating Ontologies with Large Language Models for Enhanced Control Systems in Chemical Engineering
• The Quest for Generalizable Motion Generation: Data, Model, and Evaluation
• An All-Reduce Compatible Top-K Compressor for Communication-Efficient Distributed Learning
• Encoder-Decoder or Decoder-Only? Revisiting Encoder-Decoder Large Language Model
• SecureReviewer: Enhancing Large Language Models for Secure Code Review through Secure-aware Fine-tuning
• 1+1>2: A Synergistic Sparse and Low-Rank Compression Method for Large Language Models
• UniTok-Audio: A Unified Audio Generation Framework via Generative Modeling on Discrete Codec Tokens
• Don't Let It Fade: Preserving Edits in Diffusion Language Models via Token Timestep Allocation
• Bridging the Gap Between Molecule and Textual Descriptions via Substructure-aware Alignment
• Towards Scaling Laws for Symbolic Regression
• Predicate Renaming via Large Language Models
• NeuronMM: High-Performance Matrix Multiplication for LLM Inference on AWS Trainium
• Revisiting Multilingual Data Mixtures in Language Model Pretraining
• SoK: Honeypots & LLMs, More Than the Sum of Their Parts?
• VFXMaster: Unlocking Dynamic Visual Effect Generation via In-Context Learning
• Benchmarking Generative AI Against Bayesian Optimization for Constrained Multi-Objective Inverse Design
• How Small Can You Go? A Controlled Study of LoRA Rank, Target Modules, and Quantization Trade-offs for Text-to-SQL on a 60M-Parameter Model
• ClockRoPE: Random Fourier Rotations for Temporal Routine Modeling
• Explicit Note-Event Tokenization and Pitch-Validity Constrained Decoding for MIDI-to-Tablature Transcription
• CG-World: A Large-Scale World-State Dataset and Protocol for World Models
• Between Gradient and Natural Gradient: A Continuum of LoRA Initializations
• Choosing Where and How to Moderate: End-to-End Trade-offs in Filter Placement and Response Rewriting
• Mixture-of-Depths Attention
• Training Language Models via Neural Cellular Automata
• TridentServe: A Stage-level Serving System for Diffusion Pipelines
• Taming Text-to-Sounding Video Generation via Advanced Modality Condition and Interaction
• Truth-Aware Decoding: A Program-Logic Approach to Factual Language Generation
• MedFusionT5: Cross-Modal Attention Boosts Semantic Quality and Reduces Hallucinations in Dental AI
• The Scaling Properties of Implicit Deductive Reasoning in Transformers
• MANDO-LLM: Heterogeneous Graph Transformers with Large Language Models for Smart Contract Vulnerability Detection
• HRM-Text: Efficient Pretraining Beyond Scaling
• Understanding Wacky Weights: A Dissection of SPLADE's Learned Term Importance
• A Bitter Lesson for Data Filtering
• CatPath‐GPT: A Mixture of Experts System for Computational Catalyst Design
• SHIELD: Suppressing Hallucinations In LVLM Encoders via Bias and Vulnerability Defense
• On Surprising Effectiveness of Masking Updates in Adaptive Optimizers
• Enhancing Compositional Reasoning in CLIP via Reconstruction and Alignment of Text Descriptions
• RL makes MLLMs see better than SFT
• Probing the Hidden Talent of ASR Foundation Models for L2 English Oral Assessment
• Beyond One-Size-Fits-All: Personalized Harmful Content Detection with In-Context Learning
• IBNorm: Information-Bottleneck Inspired Normalization for Representation Learning
• PSTF-AttControl: Per-Subject-Tuning-Free Personalized Image Generation with Controllable Face Attributes
• Sequences of Logits Reveal the Low Rank Structure of Language Models
• Language Model Behavioral Phases are Consistent Across Architecture, Training Data, and Scale
• Iterative Critique-Refine Framework for Enhancing LLM Personalization
• Group Relative Attention Guidance for Image Editing
• MISA: Memory-Efficient LLMs Optimization with Module-wise Importance Sampling
• From Cross-Task Examples to In-Task Prompts: A Graph-Based Pseudo-Labeling Framework for In-context Learning
• REALM: An MLLM-Agent Framework for Open World 3D Reasoning Segmentation and Editing on Gaussian Splatting
• TokenAR: Multiple Subject Generation via Autoregressive Token-level enhancement
• Exploring the Influence of Relevant Knowledge for Natural Language Generation Interpretability
• MeCeFO: Enhancing LLM Training Robustness via Fault-Tolerant Optimization
• Utilising Large Language Models for Generating Effective Counter Arguments to Anti-Vaccine Tweets
• Diffusion Adaptive Text Embedding for Text-to-Image Diffusion Models
• Improving the Straight-Through Estimator with Zeroth-Order Information
• SSL4RL: Revisiting Self-supervised Learning as Intrinsic Reward for Visual-Language Reasoning
• MoS-VLA: A Vision-Language-Action Model with One-Shot Skill Adaptation
• Modeling Expert Interactions in Sparse Mixture of Experts via Graph Structures
• More Than Generation: Unifying Generation and Depth Estimation via Text-to-Image Diffusion Models
• Large language model-based task planning for service robots: A review
• Autoregressive Styled Text Image Generation, but Make it Reliable
• Seeing Through the Brain: New Insights from Decoding Visual Stimuli with fMRI
• SwiftTS: A Swift Selection Framework for Time Series Pre-trained Models via Multi-task Meta-Learning
• Encoder-Decoder Diffusion Language Models for Efficient Training and Inference
• SeeDNorm: Self-Rescaled Dynamic Normalization
• Multi-Modal Fact-Verification Framework for Reducing Hallucinations in Large Language Models
• SALSA: Single-pass Autoregressive LLM Structured Classification
• Frustratingly Easy Task-aware Pruning for Large Language Models
• E2Rank: Your Text Embedding can Also be an Effective and Efficient Listwise Reranker
• The Structural Scalpel: Automated Contiguous Layer Pruning for Large Language Models
• Expert Merging in Sparse Mixture of Experts with Nash Bargaining
• PromptReverb: Multimodal Room Impulse Response Generation Through Latent Rectified Flow Matching
• NetBurst: Event-Centric Forecasting of Bursty, Intermittent Time Series
• Label Smoothing Improves Gradient Ascent in LLM Unlearning
• HARMONY: Hidden Activation Representations and Model Output-Aware Uncertainty Estimation for Vision-Language Models
• SentiMaithili: A Benchmark Dataset for Sentiment and Reason Generation for the Low-Resource Maithili Language
• Foundation of Intelligence: Review of Math Word Problems from Human Cognition Perspective
• Foley Control: Aligning a Frozen Latent Text-to-Audio Model to Video
• PARL: Prompt-based Agents for Reinforcement Learning
• Flight Delay Prediction via Cross-Modality Adaptation of Large Language Models and Aircraft Trajectory Representation
• FairImagen: Post-Processing for Bias Mitigation in Text-to-Image Models
• Chronos-2: From Univariate to Universal Forecasting
• Bi-Level Optimization for Generative Recommendation: Bridging Tokenization and Generation
• A Reinforcement Learning Framework for Robust and Secure LLM Watermarking
• xMem: A CPU-Based Approach for Accurate Estimation of GPU Memory in Deep Learning Training Workloads
• Learning Grouped Lattice Vector Quantizers for Low-Bit LLM Compression
• Video Prediction of Dynamic Physical Simulations With Pixel-Space Spatiotemporal Transformers
• Plan Then Retrieve: Reinforcement Learning-Guided Complex Reasoning over Knowledge Graphs
• Improving named entity correctness of abstractive summarization by generative negative sampling
• Layer as Puzzle Pieces: Compressing Large Language Models through Layer Concatenation
• Evaluating Latent Knowledge of Public Tabular Datasets in Large Language Models
• RECALL: REpresentation-aligned Catastrophic-forgetting ALLeviation via Hierarchical Model Merging
• From Masks to Worlds: A Hitchhiker's Guide to World Models
• Relative-Based Scaling Law for Neural Language Models
• Better Tokens for Better 3D: Advancing Vision-Language Modeling in 3D Medical Imaging
• A Concrete Roadmap towards Safety Cases based on Chain-of-Thought Monitoring
• Data-Centric Lessons To Improve Speech-Language Pretraining
• Fast Inference via Hierarchical Speculative Decoding
• Energy-Efficient and Dequantization-Free Q-LLMs: A Spiking Neural Network Approach to Salient Value Mitigation
• Machine Text Detectors are Membership Inference Attacks
• ELUTQ: Efficient LUT-Aware Quantization for Deploying Large Language Models on Edge Devices
• Restoring Pruned Large Language Models via Lost Component Compensation
• CPSVD: Enhancing Large Language Model Compression via Column-Preserving Singular Value Decomposition
• Difficulty-Controllable Multiple-Choice Question Generation Using Large Language Models and Direct Preference Optimization
• Selecting and Combining Large Language Models for Scalable Code Clone Detection
• ARA: Adaptive Rank Allocation for Efficient Large Language Model SVD Compression
• CAGE: Curvature-Aware Gradient Estimation For Accurate Quantization-Aware Training
• Reasoning Language Model Inference Serving Unveiled: An Empirical Study
• Large language models for folktale type automation based on motifs: Cinderella case study
• Large-scale User Game Lifecycle Representation Learning
• From Retrieval to Generation: Unifying External and Parametric Knowledge for Medical Question Answering
• Learning from the Best, Differently: A Diversity-Driven Rethinking on Data Selection
• ssToken: Self-modulated and Semantic-aware Token Selection for LLM Fine-tuning
• CircuitSeer: Mining High-Quality Data by Probing Mathematical Reasoning Circuits in LLMs
• Identity-Aware Large Language Models require Cultural Reasoning
• Towards Faithful and Controllable Personalization via Critique-Post-Edit Reinforcement Learning
• Benchmarking Probabilistic Time Series Forecasting Models on Neural Activity
• Unbiased Gradient Low-Rank Projection
• Foundational Automatic Evaluators: Scaling Multi-Task Generative Evaluator Training for Reasoning-Centric Domains
• Contextual Attention Modulation: Towards Efficient Multi-Task Adaptation in Large Language Models
• DETree: DEtecting Human-AI Collaborative Texts via Tree-Structured Hierarchical Representation Learning
• AFRICAPTION: Establishing a New Paradigm for Image Captioning in African Languages
• DSEBench: A Test Collection for Explainable Dataset Search with Examples
• Rethinking On-policy Optimization for Query Augmentation
• AION-1: Omnimodal Foundation Model for Astronomical Sciences
• Generation then Reconstruction: Accelerating Masked Autoregressive Models via Two-Stage Sampling
• Annotation-Efficient Universal Honesty Alignment
• Watermark Robustness and Radioactivity May Be at Odds in Federated Learning
• Online Learning Defense against Iterative Jailbreak Attacks via Prompt Optimization
• Back to Bytes: Revisiting Tokenization Through UTF-8
• Connecting Domains and Contrasting Samples: A Ladder for Domain Generalization
• Uncovering Brain-Like Hierarchical Patterns in Vision-Language Models through fMRI-Based Neural Encoding
• Neuronal Group Communication for Efficient Neural representation
• All You Need is One: Capsule Prompt Tuning with a Single Vector
• Improving Model Representation and Reducing KV Cache via Skip Connections with First Value Heads
• Mixed-Precision Quantization for Language Models: Techniques and Prospects
• Safe and Efficient In-Context Learning via Risk Control
• Multi-dimensional Data Analysis and Applications Basing on LLM Agents and Knowledge Graph Interactions
• Beyond Accuracy: Are Time Series Foundation Models Well-Calibrated?
• Extending Audio Context for Long-Form Understanding in Large Audio-Language Models
• SNOO: Step-K Nesterov Outer Optimizer - The Surprising Effectiveness of Nesterov Momentum Applied to Pseudo-Gradients
• Policy Transfer for Continuous-Time Reinforcement Learning: A (Rough) Differential Equation Approach
• FarsiMCQGen: a Persian Multiple-choice Question Generation Framework
• DMRetriever: A Family of Models for Improved Text Retrieval in Disaster Management
• DeLeaker: Dynamic Inference-Time Reweighting For Semantic Leakage Mitigation in Text-to-Image Models
• Midtraining Bridges Pretraining and Posttraining Distributions
• ImagerySearch: Adaptive Test-Time Search for Video Generation Beyond Semantic Dependency Constraints
• FraQAT: Quantization Aware Training with Fractional bits
• Seesaw: Accelerating Training by Balancing Learning Rate and Batch Size Scheduling
• Efficient Seq2seq Coreference Resolution Using Entity Representations
• Hierarchical Semantic Retrieval with Cobweb
• Hi-Agent: Hierarchical Vision-Language Agents for Mobile Device Control
• DOS: Directional Object Separation in Text Embeddings for Multi-Object Image Generation
• Large Reasoning Embedding Models: Towards Next-Generation Dense Retrieval Paradigm
• Constraint-Driven Small Language Models Based on Agent and OpenAlex Knowledge Graph: Mining Conceptual Pathways and Discovering Innovation Points in Academic Papers
• LiteStage: Latency-aware Layer Skipping for Multi-stage Reasoning
• Inferred global dense residue transition graphs from primary structure sequences enable protein interaction prediction via directed graph convolutional neural networks
• Nondeterminism-Aware Optimistic Verification for Floating-Point Neural Networks
• FedHFT: Efficient Federated Finetuning with Heterogeneous Edge Clients
• REAP the Experts: Why Pruning Prevails for One-Shot MoE compression
• The German Commons - 154 Billion Tokens of Openly Licensed Text for German Language Models
• LTR-ICD: A Learning-to-Rank Approach for Automatic ICD Coding
• PhysMaster: Mastering Physical Representation for Video Generation via Reinforcement Learning
• Program of Thoughts for Financial Reasoning: Leveraging Dynamic In-Context Examples and Generative Retrieval
• Hypernetworks for Perspectivist Adaptation
• Don't Be Greedy, Just Relax! Pruning LLMs via Frank-Wolfe
• Narrow Finetuning Leaves Clearly Readable Traces in Activation Differences
• Cautious Weight Decay
• Refine Thought: A Test-Time Inference Method for Embedding Model Reasoning
• Reinforced Preference Optimization for Recommendation
• Chimera: State Space Models Beyond Sequences
• Ethic-BERT: An Enhanced Deep Learning Model for Ethical and Non-Ethical Content Classification
• Readout Representation: Redefining Neural Codes by Input Recovery
• Catch Your Breath: Adaptive Computation for Self-Paced Sequence Production
• Point Prompting: Counterfactual Tracking with Video Diffusion Models
• GenCNER: A Generative Framework for Continual Named Entity Recognition
• Are Large Language Models Effective Knowledge Graph Constructors?
• Vision-LLMs for Spatiotemporal Traffic Forecasting
• Neural Weight Compression for Language Models
• A Theorem-Proving-Based Evaluation of Neural Semantic Parsing
• Domain-Specific Data Generation Framework for RAG Adaptation
• An Explorative Study on Distributed Computing Techniques in Training and Inference of Large Language Models
• Class Prototypes based Contrastive Learning for Classifying Multi-Label and Fine-Grained Educational Videos
• Evaluating Line-level Localization Ability of Learning-based Code Vulnerability Detection Models
• Saudi Sign Language Translation Using T5
• CoSPED: Consistent Soft Prompt Targeted Data Extraction and Defense
• MC#: Mixture Compressor for Mixture-of-Experts Large Models
• Learning to Watermark: A Selective Watermarking Framework for Large Language Models via Multi-Objective Optimization
• PAGE: Prompt Augmentation for text Generation Enhancement
• Direct Multi-Token Decoding
• MatterDoor: Sampling Zero-shot Spatio-semantic Priors using Generative Models
• Early Detection and Reduction of Memorisation for Domain Adaptation and Instruction Tuning
• Preserving LLM Capabilities through Calibration Data Curation: From Analysis to Optimization
• AnyBCQ: Hardware Efficient Flexible Binary-Coded Quantization for Multi-Precision LLMs
• Head-wise Adaptive Rotary Positional Encoding for Fine-Grained Image Generation
• Rethinking LLM Evaluation: Can We Evaluate LLMs with 200x Less Data?
• R2T: Rule-Encoded Loss Functions for Low-Resource Sequence Tagging
• UltraLLaDA: Scaling the Context Length to 128K for Diffusion Large Language Models
• ReMix: Towards a Unified View of Consistent Character Generation and Editing
• PermLLM: Learnable Channel Permutation for N:M Sparse Large Language Models
• Translution: Unifying Self-attention and Convolution for Adaptive and Relative Modeling
• Small is Sufficient: Reducing the World AI Energy Consumption Through Model Selection
• Forget Attention: Importance-Aware Attention Is All You Need
• SCRIBES: Web-Scale Script-Based Semi-Structured Data Extraction with Reinforcement Learning
• Sparse Query Attention (SQA): A Computationally Efficient Attention Mechanism with Query Heads Reduction
• PyramidStyler: Transformer-Based Neural Style Transfer with Pyramidal Positional Encoding and Reinforcement Learning
• Lost in the Middle: An Emergent Property from Information Retrieval Demands in LLMs
• PatentVision: A multimodal method for drafting patent applications
• Patentformer: A demonstration of AI-assisted automated patent drafting
• Hybrid Models for Natural Language Reasoning: The Case of Syllogistic Logic
• The Speech-LLM Takes It All: A Truly Fully End-to-End Spoken Dialogue State Tracking Approach
• AdaPM: a Partial Momentum Algorithm for LLM Training
• Provable Watermarking for Data Poisoning Attacks
• Web Crawler Restrictions, AI Training Datasets & Political Biases
• The Potential of Second-Order Optimization for LLMs: A Study with Full Gauss-Newton
• Diffusion-Inspired Masked Fine-Tuning for Knowledge Injection in Autoregressive LLMs
• Q-Router: Agentic Video Quality Assessment with Expert Model Routing and Artifact Localization
• UniVideo: Unified Understanding, Generation, and Editing for Videos
• UniMMVSR: A Unified Multi-Modal Framework for Cascaded Video Super-Resolution
• Fewer Weights, More Problems: A Practical Attack on LLM Pruning
• Vocabulary embeddings organize linguistic structure early in language model training
• LOTION: Smoothing the Optimization Landscape for Quantized Training
• DACIP-RC: Domain Adaptive Continual Instruction Pre-Training via Reading Comprehension on Business Conversations
• Multi-Task Pre-Finetuning of Lightweight Transformer Encoders for Text Classification and NER
• On the Convergence of Moral Self-Correction in Large Language Models
• Sunflower: A New Approach To Expanding Coverage of African Languages in Large Language Models
• Encode, Think, Decode: Scaling test-time reasoning with recursive latent thoughts
• Vision-Language-Action Models for Robotics: A Review Towards Real-World Applications
• Textual interpretation of transient image classifications from large language models
• SoftMatcha 2: A Fast and Soft Pattern Matcher for Trillion-Scale Corpora
• Vectorized FlashAttention with Low-cost Exponential Computation in RISC-V Vector Processors
• Mid-Training of Large Language Models: A Survey
• TalkCuts: A Large-Scale Dataset for Multi-Shot Human Speech Video Generation
• GUIDE: Guided Initialization and Distillation of Embeddings
• Webscale-RL: Automated Data Pipeline for Scaling RL Data to Pretraining Levels
• Test-Time Efficient Pretrained Model Portfolios for Time Series Forecasting
• Training Dynamics Impact Post-Training Quantization Robustness
• Gradient-Sign Masking for Task Vector Transport Across Pre-Trained Models
• LARA-Gen: Enabling Continuous Emotion Control for Music Generation Models via Latent Affective Representation Alignment
• Mixture of Neuron Experts
• Membership Inference Attacks on Tokenizers of Large Language Models
• Paraplume: A fast and accurate paratope prediction method provides insights into repertoire-scale binding dynamics
• BLISS: A Lightweight Bilevel Influence Scoring Method for Data Selection in Language Model Pretraining
• Data Provenance Auditing of Fine-Tuned Large Language Models with a Text-Preserving Technique
• Factuality Matters: When Image Generation and Editing Meet Structured Visuals
• SAEdit: Token-level control for continuous image editing via Sparse AutoEncoder
• Adaptive Memory Momentum via a Model-Based Framework for Deep Learning Optimization
• HyperVLA: Efficient Inference in Vision-Language-Action Models via Hypernetworks
• Demystifying Synthetic Data in LLM Pre-training: A Systematic Study of Scaling Laws, Benefits, and Pitfalls
• SpikingMamba: Towards Energy-Efficient Large Language Models via Knowledge Distillation from Mamba
• Representation Potentials of Foundation Models for Multimodal Alignment: A Survey
• UniVoice: Unifying Autoregressive ASR and Flow-Matching based TTS with Large Language Models
• Inoculation Prompting: Eliciting traits from LLMs during training can suppress them at test-time
• FoilDiff: A Hybrid Transformer Backbone for Diffusion-based Modelling of 2D Airfoil Flow Fields
• Activation Steering with a Feedback Controller
• Thai Semantic End-of-Turn Detection for Real-Time Voice Agents
• Large Language Models Hallucination: A Comprehensive Survey
• Spectral Alignment as Predictor of Loss Explosion in Neural Network Training
• The Unseen Frontier: Pushing the Limits of LLM Sparsity with Surrogate-Free ADMM
• Smart Paste: Automatically Fixing Copy/Paste for Google Developers
• Allocation of Parameters in Transformers
• On the Empirical Power of Goodness-of-Fit Tests in Watermark Detection
• CarbonX: An Open-Source Tool for Computational Decarbonization Using Time Series Foundation Models
• When and Where do Events Switch in Multi-Event Video Generation?
• DMark: Order-Agnostic Watermarking for Diffusion Large Language Models
• Distributed Low-Communication Training with Decoupled Momentum Optimization
• MoGIC: Boosting Motion Generation via Intention Understanding and Visual Context
• AgenticRAG: Tool-Augmented Foundation Models for Zero-Shot Explainable Recommender Systems
• Brain-Language Model Alignment: Insights into the Platonic Hypothesis and Intermediate-Layer Advantage
• Uncovering the Computational Ingredients of Human-Like Representations in LLMs
• Benchmarking Foundation Models with Retrieval-Augmented Generation in Olympic-Level Physics Problem Solving
• HalluGuard: Evidence-Grounded Small Reasoning Models to Mitigate Hallucinations in Retrieval-Augmented Generation
• Erased, But Not Forgotten: Erased Rectified Flow Transformers Still Remain Unsafe Under Concept Attack
• CML-Bench: A Framework for Evaluating and Enhancing LLM-Powered Movie Scripts Generation
• Black-Box Time-Series Domain Adaptation via Cross-Prompt Foundation Models
• Microsaccade-Inspired Probing: Positional Encoding Perturbations Reveal LLM Misbehaviours
• Decomposing Attention To Find Context-Sensitive Neurons
• Fast, Secure, and High-Capacity Image Watermarking with Autoencoded Text Vectors
• Retrieval-Augmented Framework for LLM-Based Clinical Decision Support
• Learn to Guide Your Diffusion Model
• BindWeave: Subject-Consistent Video Generation via Cross-Modal Integration
• Train Large, Deploy Compact: Structured Compression for Compact Low-Rank Adaptation
• Stitch: Training-Free Position Control in Multimodal Diffusion Transformers
• Probing the Critical Point (CritPt) of AI Reasoning: a Frontier Physics Research Benchmark
• Bayesian Influence Functions for Hessian-Free Data Attribution
• Regression Language Models for Code
• Adaptive Planning for Multi-Attribute Controllable Summarization with Monte Carlo Tree Search
• IMG: Calibrating Diffusion Models via Implicit Multimodal Guidance
• PatchVSR: Breaking Video Diffusion Resolution Limits with Patch-wise Video Super-Resolution
• Indirect Attention: Turning Context Misalignment into a Feature
• Understanding the Mixture-of-Experts with Nadaraya-Watson Kernel
• Revoking Amnesia: RL-based Trajectory Optimization to Resurrect Erased Concepts in Diffusion Models
• From Cheap Geometry to Expensive Physics: Elevating Neural Operators via Latent Shape Pretraining
• Layer-wise dynamic rank for compressing large language models
• Per-example gradients: a new frontier for understanding and improving optimizers
• MixtureVitae: Open Web-Scale Pretraining Dataset With High Quality Instruction and Reasoning Data Built from Permissive-First Text Sources
• Scaling with Collapse: Efficient and Predictable Training of LLM Families
• SAGA-SR: Semantically and Acoustically Guided Audio Super-Resolution
• From Code to Action: Hierarchical Learning of Diffusion-VLM Policies
• Pushing LLMs to Their Logical Reasoning Bound: The Role of Data Reasoning Intensity
• UniPruning: Unifying Local Metric and Global Feedback for Scalable Sparse LLMs
• Intent-Driven Storage Systems: From Low-Level Tuning to High-Level Understanding
• Can you SPLICE it together? A Human Curated Benchmark for Probing Visual Reasoning in VLMs
• AdaDetectGPT: Adaptive Detection of LLM-Generated Text with Statistical Guarantees
• Multilingual Text-to-SQL: Benchmarking the Limits of Language Models with Collaborative Language Agents
• Conda: Column-Normalized Adam for Training Large Language Models Faster
• PET: Preference Evolution Tracking with LLM-Generated Explainable Distribution
• Watermarking Diffusion Language Models
• PanoWorld-X: Generating Explorable Panoramic Worlds via Sphere-Aware Video Diffusion
• AGNOMIN -- Architecture Agnostic Multi-Label Function Name Prediction
• <scp>AI</scp> Methods for Antimicrobial Peptides: Progress and Challenges
• MapGenerator: a framework for learning a diffusion model for text promptable map generation
• The Rise of AfricaNLP: A Survey of Contributions, Contributors, Community Impact, and Bibliometric Analysis
• Reinforcement Mid-Training
• Analyzing and Evaluating Unbiased Language Model Watermark
• An Ensemble Framework for Unbiased Language Model Watermarking
• Dynamic Orthogonal Continual Fine-tuning for Mitigating Catastrophic Forgettings
• Investigating Multi-layer Representations for Dense Passage Retrieval
• GSID: Generative Semantic Indexing for E-Commerce Product Understanding
• Estimating Time Series Foundation Model Transferability via In-Context Learning
• MotionVerse: A Unified Multimodal Framework for Motion Comprehension, Generation and Editing
• Training Optimal Large Diffusion Language Models
• Temporal Generalization: A Reality Check
• A systematic evaluation of Dutch large language models’ surprisal estimates in sentence, paragraph and book reading
• WorldSplat: Gaussian-Centric Feed-Forward 4D Scene Generation for Autonomous Driving
• SDQ-LLM: Sigma-Delta Quantization for 1-bit LLMs of any size
• An Improved Framework for Scaling Party Positions from Texts with Transformer
• Fact Grounded Attention: Eliminating Hallucination in Large Language Models Through Attention Level Knowledge Integration
• PDE-Transformer: A Continuous Dynamical Systems Approach to Sequence Modeling
• How to Make Large Language Models Generate 100% Valid Molecules?
• PT²-LLM: Post-Training Ternarization for Large Language Models
• PlasGO: enhancing GO-based function prediction for plasmid-encoded proteins based on genetic structure
• LLM Watermark Evasion via Bias Inversion
• SINQ: Sinkhorn-Normalized Quantization for Calibration-Free Low-Precision LLM Weights
• WoW: Towards a World omniscient World model Through Embodied Interaction
• JointDiff: Bridging Continuous and Discrete in Multi-Agent Trajectory Generation
• A model of errors in transformers
• What Is The Political Content in LLMs' Pre- and Post-Training Data?
• Advancing Natural Language Formalization to First Order Logic with Fine-tuned LLMs
• Beyond Textual Context: Structural Graph Encoding with Adaptive Space Alignment to alleviate the hallucination of LLMs
• Context Parametrization with Compositional Adapters
• Log2Plan: An Adaptive GUI Automation Framework Integrated with Task Mining Approach
• Multilingual Vision-Language Models, A Survey
• FailureAtlas:Mapping the Failure Landscape of T2I Models via Active Exploration
• Syncphony: Synchronized Audio-to-Video Generation with Diffusion Transformers
• Enhancing Low-Rank Adaptation with Structured Nonlinear Transformations
• Bridging Kolmogorov Complexity and Deep Learning: Asymptotically Optimal Description Length Objectives for Transformers
• EMMA: Generalizing Real-World Robot Manipulation via Generative Visual Transfer
• RLP: Reinforcement as a Pretraining Objective
• Towards Transparent AI: A Survey on Explainable Language Models
• SD3.5-Flash: Distribution-Guided Distillation of Generative Flows
• Query-Centric Graph Retrieval Augmented Generation
• IntSR: An Integrated Generative Framework for Search and Recommendation
• Distributed Specialization: Rare-Token Neurons in Large Language Models
• PMark: Towards Robust and Distortion-free Semantic-level Watermarking with Channel Constraints
• A short survey on almost orthogonal vectors in a few specific large dimensions
• RLCracker: Evaluating the Worst-Case Vulnerability of LLM Watermarks with Adaptive RL Attacks
• FORGE: Forming Semantic Identifiers for Generative Retrieval in Industrial Datasets
• Prompt-Aware Scheduling for Low-Latency LLM Serving
• Zero-Shot Privacy-Aware Text Rewriting via Iterative Tree Search
• Leveraging What's Overfixed: Post-Correction via LLM Grammatical Error Overcorrection
• Measuring LLM Sensitivity in Transformer-based Tabular Data Synthesis
• Overcoming Black-box Attack Inefficiency with Hybrid and Dynamic Select Algorithms
• Concise and Sufficient Sub-Sentence Citations for Retrieval-Augmented Generation
• RedHerring Attack: Testing the Reliability of Attack Detection
• Unlocking Financial Insights: An advanced Multimodal Summarization with Multimodal Output Framework for Financial Advisory Videos
• Understanding and Enhancing Mask-Based Pretraining towards Universal Representations
• Performance Consistency of Learning Methods for Information Retrieval Tasks
• OLaPh: Optimal Language Phonemizer
• Embodied AI: From LLMs to World Models
• Faster, Smaller, and Smarter: Task-Aware Expert Merging for Online MoE Inference
• Learning Contextual Retrieval for Robust Conversational Search
• SeMob: Semantic Synthesis for Dynamic Urban Mobility Prediction
• Mamba Modulation: On the Length Generalization of Mamba
• ExPe: Exact Positional Encodings for Generative Transformer Models with Extrapolating Capabilities
• Looped Transformers with Source-Centered State Evolution
• ReGenVC: End-to-End Real-Time Generative Video Coding at Ultra-Low Bitrate
• What Comes Next? Evaluating Uncertainty in Neural Text Generators Against Human Production Variability
• Hand-Object Interaction in the Age of Large Foundation Models:Reconstruction, Generation, and Embodied Transfer
• OneShot: Index-in-Ranking with Neural Scoring for Large-Scale Retrieval
• CodeBERT: A Pre-Trained Model for Programming and Natural Languages
• BridgeAlign: Bridging Preference Alignment for Humanities and Social Sciences
• MAD: Motion Appearance Decoupling for efficient Driving World Models
• HSS-Synth: Humanities and Social Sciences Data Synthesis for LLMs
• MedLLM: An Open Medical Language Model at the Sub-Billion Scale
• Averaged Evaluation Masks Capability Trade-Offs: Multi-Source Calibration for High-Sparsity LLM Pruning
• ELF: Embedded Language Flows
• How Can We Synthesize High-Quality Pretraining Data? A Systematic Study of Prompt Design, Generator Model, and Source Data
• RELISH: LLM REgression with a Latent Iterative State Head
• Legal text summarization via judicial syllogism with large language models
• ChatGPT: More Than a “Weapon of Mass Deception” Ethical Challenges and Responses from the Human-Centered Artificial Intelligence (HCAI) Perspective
• GradMAP: Faster Layer Pruning with Gradient Metric and Projection Compensation
• Transporting Task Vectors across Different Architectures without Training
• BaGuaLu
• epiGPTope: A Machine Learning-Based Epitope Generator and Classifier
• AI for social science and social science of AI: A survey
• CamemBERT: a Tasty French Language Model
• LLMulator: Generalizable Cost Modeling for Dataflow Accelerators with Input-Adaptive Control Flow
• Speculating LLMs' Chinese Training Data Pollution from Their Tokens
• Randomly Removing 50% of Dimensions in Text Embeddings has Minimal Impact on Retrieval and Classification Tasks
• Reinforcement Learning on Pre-Training Data
• Pure Vision Language Action (VLA) Models: A Comprehensive Survey
• CR-Net: Scaling Parameter-Efficient Training with Cross-Layer Low-Rank Structure
• Memory in Large Language Models: Mechanisms, Evaluation and Evolution
• Text Slider: Efficient and Plug-and-Play Continuous Concept Control for Image/Video Synthesis via LoRA Adapters
• AGSwap: Overcoming Category Boundaries in Object Fusion via Adaptive Group Swapping
• Prior-based Noisy Text Data Filtering: Fast and Strong Alternative For Perplexity
• Trigger Where It Hurts: Unveiling Hidden Backdoors through Sensitivity with Sensitron
• False Friends Are Not Foes: Investigating Vocabulary Overlap in Multilingual Language Models
• Financial Risk Relation Identification through Dual-view Adaptation
• OmniBridge: Unified Multimodal Understanding, Generation, and Retrieval via Latent Space Alignment
• Data Efficient Adaptation in Large Language Models via Continuous Low-Rank Fine-Tuning
• A Rhythm-Aware Phrase Insertion for Classical Arabic Poetry Composition
• An overview of neural architectures for self-supervised audio representation learning from masked spectrograms
• Advances in Large Language Models for Medicine
• Enhancing Transformer-Based Rerankers with Synthetic Data and LLM-Based Supervision
• Actions Speak Louder than Prompts: A Large-Scale Study of LLMs for Graph Inference
• Are Smaller Open-Weight LLMs Closing the Gap to Proprietary Models for Biomedical Question Answering?
• MOMEMTO: Patch-based Memory Gate Model in Time Series Foundation Model
• Seg4Diff: Unveiling Open-Vocabulary Segmentation in Text-to-Image Diffusion Transformers
• The Narcissus Hypothesis: Descending to the Rung of Illusion
• Transformer-Encoder Trees for Efficient Multilingual Machine Translation and Speech Translation
• CorefInst: Leveraging LLMs for Multilingual Coreference Resolution
• UIPro: Unleashing Superior Interaction Capability For GUI Agents
• OpenGVL -- Benchmarking Visual Temporal Progress for Data Curation
• AIMMerging: Adaptive Iterative Model Merging Using Training Trajectories for Language Model Continual Learning
• PTQTP: Post-Training Quantization to Trit-Planes for Large Language Models
• Variational Task Vector Composition
• Analyzing Memory Effects in Large Language Models through the lens of Cognitive Psychology
• MCP: A Control-Theoretic Orchestration Framework for Synergistic Efficiency and Interpretability in Multimodal Large Language Models
• Less Is More? Examining Fairness in Pruned Large Language Models for Summarising Opinions
• The Oracle Has Spoken: A Multi-Aspect Evaluation of Dialogue in Pythia
• A Multi-Level Benchmark for Causal Language Understanding in Social Media Discourse
• Text-Scene: A Scene-to-Language Parsing Framework for 3D Scene Understanding
• TinyGiantVLM: A Lightweight Vision-Language Architecture for Spatial Reasoning under Resource Constraints
• A Unified AI Approach for Continuous Monitoring of Human Health and Diseases from Intensive Care Unit to Home with Physiological Foundation Models (UNIPHY+)
• DiEP: Adaptive Mixture-of-Experts Compression through Differentiable Expert Pruning
• Vision-Language Models as Differentiable Semantic and Spatial Rewards for Text-to-3D Generation
• DNA-DetectLLM: Unveiling AI-Generated Text via a DNA-Inspired Mutation-Repair Paradigm
• Purely Semantic Indexing for LLM-based Generative Recommendation and Retrieval
• Hierarchical Retrieval: The Geometry and a Pretrain-Finetune Recipe
• Chunk Knowledge Generation Model for Enhanced Information Retrieval: A Multi-task Learning Approach
• Structured Information for Improving Spatial Relationships in Text-to-Image Generation
• Evaluating the Effectiveness and Scalability of LLM-Based Data Augmentation for Retrieval
• On the Convergence of Muon and Beyond
• Generalizability of Large Language Model-Based Agents: A Comprehensive Survey
• Efficient and Versatile Model for Multilingual Information Retrieval of Islamic Text: Development and Deployment in Real-World Scenarios
• Understand Before You Generate: Self-Guided Training for Autoregressive Image Generation
• Language Modeling with Learned Meta-Tokens
• A Comparative Analysis of Transformer Models in Social Bot Detection
• Evaluating the Impact of Verbal Multiword Expressions on Machine Translation
• MaiBERT: A Pre-training Corpus and Language Model for Low-Resourced Maithili Language
• Super-Linear: A Lightweight Pretrained Mixture of Linear Experts for Time Series Forecasting
• Deep learning and abstractive summarisation for radiological reports: an empirical study for adapting the PEGASUS models' family with scarce data
• Hierarchical Self-Attention: Generalizing Neural Attention Mechanics to Multi-Scale Problems
• Fair-GPTQ: Bias-Aware Quantization for Large Language Models
• Synthetic bootstrapped pretraining
• A Framework for Generating Artificial Datasets to Validate Absolute and Relative Position Concepts
• MoSA: Motion-Coherent Human Video Generation via Structure-Appearance Decoupling
• Wan-Animate: Unified Character Animation and Replacement with Holistic Replication
• Enhancing Time Awareness in Generative Recommendation
• TFMAdapter: Lightweight Instance-Level Adaptation of Foundation Models for Forecasting with Covariates
• Condition Weaving Meets Expert Modulation: Towards Universal and Controllable Image Generation
• How Can Quantum Deep Learning Improve Large Language Models?
• Dual-Actor Fine-Tuning of VLA Models: A Talk-and-Tweak Human-in-the-Loop Approach
• TENET: An Efficient Sparsity-Aware LUT-Centric Architecture for Ternary LLM Inference On Edge
• Re-purposing SAM into Efficient Visual Projectors for MLLM-Based Referring Image Segmentation
• Do Activation Verbalization Methods Convey Privileged Information?
• Membership Inference Attack against Large Language Model-based Recommendation Systems: A New Distillation-based Paradigm
• Harnessing the Power of AI in Qualitative Research: Role Assignment, Engagement, and User Perceptions of AI-Generated Follow-Up Questions in Semi-Structured Interviews
• Don't Change My View: Ideological Bias Auditing in Large Language Models
• Positional Encoding via Token-Aware Phase Attention
• An Empirical Study of Retrieval-Augmented Code Generation: Challenges and Opportunities
• Towards Alignment-Centric Paradigm: A Survey of Instruction Tuning in Large Language Models
• Automated Generation of Research Workflows from Academic Papers: A Full-text Mining Framework
• Quality at a Glance: An Audit of Web-Crawled Multilingual Datasets
• Natural Language Satisfiability: Exploring the Problem Distribution and Evaluating Transformer-based Language Models
• Pun Unintended: LLMs and the Illusion of Humor Understanding
• VQL: An End-to-End Context-Aware Vector Quantization Attention for Ultra-Long User Behavior Modeling
• AMQ: Enabling AutoML for Mixed-precision Weight-Only Quantization of Large Language Models
• CoachMe: Decoding Sport Elements with a Reference-Based Coaching Instruction Generation Model
• Tenma: Robust Cross-Embodiment Robot Manipulation with Diffusion Transformer
• Reasoning Models Can be Accurately Pruned Via Chain-of-Thought Reconstruction
• DetectAnyLLM: Towards Generalizable and Robust Detection of Machine-Generated Text Across Domains and Models
• Optimal Brain Restoration for Joint Quantization and Sparsification of LLMs
• When Smiley Turns Hostile: Interpreting How Emojis Trigger LLMs' Toxicity
• Introducing Spotlight: A Novel Approach for Generating Captivating Key Information from Documents
• Judge Q: Trainable Queries for Optimized Information Retention in KV Cache Eviction
• Reasoning Under Uncertainty: Exploring Probabilistic Reasoning Capabilities of LLMs
• A Survey on Retrieval And Structuring Augmented Generation with Large Language Models
• LLM in the Middle: A Systematic Review of Threats and Mitigations to Real-World LLM-based Systems
• Contrastive Prompt Clustering for Weakly Supervised Semantic Segmentation
• Safety and Security Analysis of Large Language Models: Benchmarking Risk Profile and Harm Potential
• Interdisciplinary Research in Conversation: A Case Study in Computational Morphology for Language Documentation
• Understanding Outer Optimizers in Local SGD: Learning Rates, Momentum, and Acceleration
• DeAR: Dual-Stage Document Reranking with Reasoning Agents via LLM Distillation
• WebSight: A Vision-First Architecture for Robust Web Agents
• Dropping Experts, Recombining Neurons: Retraining-Free Pruning for Sparse Mixture-of-Experts LLMs
• Out of One, Many: Using Language Models to Simulate Human Samples
• Towards Understanding Visual Grounding in Visual Language Models
• ButterflyQuant: Ultra-low-bit LLM Quantization through Learnable Orthogonal Butterfly Transforms
• Retrieval-Augmented Generation for Reliable Interpretation of Radio Regulations
• Kling-Avatar: Grounding Multimodal Instructions for Cascaded Long-Duration Avatar Animation Synthesis
• CoPE: A Lightweight Complex Positional Encoding
• Modelling Analogies and Analogical Reasoning: Connecting Cognitive Science Theory and NLP Research
• Large Language Models in Document Intelligence: A Comprehensive Survey, Recent Advances, Challenges, and Future Trends
• Efficient Transformer-Based Piano Transcription With Sparse Attention Mechanisms
• Character-Level Perturbations Disrupt LLM Watermarks
• LLM-JEPA: Large Language Models Meet Joint Embedding Predictive Architectures
• CrossPT: Exploring Cross-Task Transferability through Multi-Task Prompt Tuning
• LLMs as Agentic Cooperative Players in Multiplayer UNO
• Artificial intelligence-driven computational methods for antibody design and optimization
• Building High-Quality Datasets for Portuguese LLMs: From Common Crawl Snapshots to Industrial-Grade Corpora
• Discrimination by LLMs: Cross-lingual Bias Assessment and Mitigation in Decision-Making and Summarisation
• Adversarial Attacks Against Automated Fact-Checking: A Survey
• Bitrate-Controlled Diffusion for Disentangling Motion and Content in Video
• Hetis: Serving LLMs in Heterogeneous GPU Clusters with Fine-grained and Dynamic Parallelism
• Open-sci-ref-0.01: open and reproducible reference baselines for language model and dataset comparison
• Handling Open-Vocabulary Constructs in Formalizing Specifications: Retrieval-Augmented Parsing with Expert Knowledge
• Guided Reasoning in LLM-Driven Penetration Testing Using Structured Attack Trees
• Query Expansion in the Age of Pre-trained and Large Language Models: A Comprehensive Survey
• Improving Machine Learning-Based Robot Self-Collision Checking with Input Positional Encoding
• Multi-view-guided Passage Reranking with Large Language Models
• Assess and Prompt: A Generative RL Framework for Improving Engagement in Online Mental Health Communities
• Dual Knowledge-Enhanced Two-Stage Reasoner for Multimodal Dialog Systems
• Mitigating Attention Localization in Small Scale: Self-Attention Refinement via One-step Belief Propagation
• CP-Model-Zoo: A Natural Language Query System for Constraint Programming Models
• From Detection to Mitigation: Addressing Gender Bias in Chinese Texts via Efficient Tuning and Voting-Based Rebalancing
• Learning Decomposed Contextual Token Representations from Pretrained and Collaborative Signals for Generative Recommendation
• COMPACT: Common-token Optimized Model Pruning Across Channels and Tokens
• TIDE: Achieving Balanced Subject-Driven Image Generation via Target-Instructed Diffusion Enhancement
• Mask-GCG: Are All Tokens in Adversarial Suffixes Necessary for Jailbreak Attacks?
• AI-driven Remote Facial Skin Hydration and TEWL Assessment from Selfie Images: A Systematic Solution
• LoaQ: Layer-wise Output Approximation Quantization
• SLiNT: Structure-aware Language Model with Injection and Contrastive Training for Knowledge Graph Completion
• HANRAG: Heuristic Accurate Noise-resistant Retrieval-Augmented Generation for Multi-hop Question Answering
• Methodological Insights into Structural Causal Modelling and Uncertainty-Aware Forecasting for Economic Indicators
• TSPC: A Two-Stage Phoneme-Centric Architecture for code-switching Vietnamese-English Speech Recognition
• Benchmarking Gender and Political Bias in Large Language Models
• DreamAudio: Customized Text-to-Audio Generation with Diffusion Models
• From Joy to Fear: A Benchmark of Emotion Estimation in Pop Song Lyrics
• Few-Shot Query Intent Detection via Relation-Aware Prompt Learning
• LatticeWorld: A Multimodal Large Language Model-Empowered Framework for Interactive Complex World Generation
• PLaMo 2 Technical Report
• L1RA: Dynamic Rank Assignment in LoRA Fine-Tuning
• Dynamic Adaptive Shared Experts with Grouped Multi-Head Attention Mixture of Experts
• Reverse Browser: Vector-Image-to-Code Generator
• Hunyuan-MT Technical Report
• Rethinking the long-range dependency in Mamba/SSM and transformer models
• SMooGPT: Stylized Motion Generation using Large Language Models
• LMAE4Eth: Generalizable and Robust Ethereum Fraud Detection by Exploring Transaction Semantics and Masked Graph Embedding
• Weakly-Supervised Learning of Dense Functional Correspondences
• Madness, Cannibalism, and Traditional Fiction between Lu Xun and Mo Yan
• Beyond Interpretability: Exploring the Comprehensibility of Adaptive Video Streaming through Large Language Models
• Skywork UniPic 2.0: Building Kontext Model with Online RL for Unified Multimodal Model
• Sali4Vid: Saliency-Aware Video Reweighting and Adaptive Caption Retrieval for Dense Video Captioning
• Cetvel: A Unified Benchmark for Evaluating Language Understanding, Generation and Cultural Capacity of LLMs for Turkish
• Binary Quantization For LLMs Through Dynamic Grouping
• Training LLMs to be Better Text Embedders through Bidirectional Reconstruction
• EverTracer: Hunting Stolen Large Language Models via Stealthy and Robust Probabilistic Fingerprint
• AIVA: An AI-based Virtual Companion for Emotion-aware Interaction
• LLMCDSR: Enhancing Cross-Domain Sequential Recommendation with Large Language Models
• Lesion-Aware Visual-Language Fusion for Automated Image Captioning of Ulcerative Colitis Endoscopic Examinations
• Implicit Reasoning in Large Language Models: A Comprehensive Survey
• Distributional Semantics: Meaning Through Culture and Interaction
• DivMerge: A divergence-based model merging method for multi-tasking
• LLMs that Understand Processes: Instruction-tuning for Semantics-Aware Process Mining
• FlexMUSE: Multimodal Unification and Semantics Enhancement Framework with Flexible interaction for Creative Writing
• Attributes as Textual Genes: Leveraging LLMs as Genetic Algorithm Simulators for Conditional Synthetic Data Generation
• MOSAIC: Multi-Subject Personalized Generation via Correspondence-Aware Alignment and Disentanglement
• Systematic Integration of Attention Modules into CNNs for Accurate and Generalizable Medical Image Diagnosis
• chDzDT: Word-level morphology-aware language model for Algerian social media text
• Clinical Metadata Guided Limited-Angle CT Image Reconstruction
• Cross-Modal Prototype Augmentation and Dual-Grained Prompt Learning for Social Media Popularity Prediction
• MixedG2P-T5: G2P-free Speech Synthesis for Mixed-script texts using Speech Self-Supervised Learning and Language Model
• Multitask Battery Management with Flexible Pretraining
• GradeSQL: Test-Time Inference with Outcome Reward Models for Text-to-SQL Generation from Large Language Models
• Towards More Diverse and Challenging Pre-training for Point Cloud Learning: Self-Supervised Cross Reconstruction with Decoupled Views
• Hierarchical Motion Captioning Utilizing External Text Data Source
• Collaborative local–global context modeling for session-based recommendation
• Two-flow Feedback Multi-scale Progressive Generative Adversarial Network
• Optimizing In-Context Learning for Efficient Full Conformal Prediction
• A Privacy-Preserving Recommender for Filling Web Forms Using a Local Large Language Model
• A case study of forensic psychiatry experts' reports analysis through large language models
• MEPT: Mixture of Expert Prompt Tuning as a Manifold Mapper
• CompSlider: Compositional Slider for Disentangled Multiple-Attribute Image Generation
• No Clustering, No Routing: How Transformers Actually Process Rare Tokens
• Mixture of Global and Local Experts with Diffusion Transformer for Controllable Face Generation
• Training Text-to-Molecule Models with Context-Aware Tokenization
• Backdoor Samples Detection Based on Perturbation Discrepancy Consistency in Pre-trained Language Models
• Not All Parameters Are Created Equal: Smart Isolation Boosts Fine-Tuning Performance
• VoCap: Video Object Captioning and Segmentation from Any Prompt
• TMUAD: Enhancing Logical Capabilities in Unified Anomaly Detection Models with a Text Memory Bank
• FLORA: Efficient Synthetic Data Generation for Object Detection in Low-Data Regimes via finetuning Flux LoRA
• QZhou-Embedding Technical Report
• Measuring Attribution in Natural Language Generation Models
• VeriLoRA: Fine-Tuning Large Language Models with Verifiable Security via Zero-Knowledge Proofs
• Going over Fine Web with a Fine-Tooth Comb: Technical Report of Indexing Fine Web for Problematic Content Search and Retrieval
• Middo: Model-Informed Dynamic Data Optimization for Enhanced LLM Fine-Tuning via Closed-Loop Learning
• Rethinking Layer-wise Model Merging through Chain of Merges
• Zero‐ and few‐shot prompting of generative large language models provides weak assessment of risk of bias in clinical trials
• Specializing General-purpose LLM Embeddings for Implicit Hate Speech Detection across Datasets
• MERIT: Maximum-normalized Element-wise Ratio for Language Model Large-batch Training
• Droplet3D: Commonsense Priors from Videos Facilitate 3D Generation
• KG-CQR: Leveraging Structured Relation Representations in Knowledge Graphs for Contextual Query Retrieval
• Waver: Wave Your Way to Lifelike Video Generation
• TCIA: A Task-Centric Instruction Augmentation Method for Instruction Finetuning
• Addressing Tokenization Inconsistency in Steganography and Watermarking Based on Large Language Models
• OneRec-V2 Technical Report
• OLMoASR: Open Models and Data for Training Robust Speech Recognition Models
• A Systematic Review on the Generative AI Applications in Human Medical Genomics
• Robustness Assessment and Enhancement of Text Watermarking for Google's SynthID
• AudioStory: Generating Long-Form Narrative Audio with Large Language Models
• Generative AI for Testing of Autonomous Driving Systems: A Survey
• TComQA: Extracting Temporal Commonsense from Text
• QuesGenie: Intelligent Multimodal Question Generation
• Bi-LoRA: Efficient Sharpness-Aware Minimization for Fine-Tuning Large-Scale Models
• ELIXIR: Efficient and LIghtweight model for eXplaIning Recommendations
• SIExVulTS: Sensitive Information Exposure Vulnerability Detection System using Transformer Models and Static Analysis
• Database Entity Recognition with Data Augmentation and Deep Learning
• MIDAS: Multimodal Interactive Digital-humAn Synthesis via Real-time Autoregressive Video Generation
• PyTOD: Programmable Task-Oriented Dialogue with Execution Feedback
• Enhancing Document VQA Models via Retrieval-Augmented Generation
• Harnessing Rule-Based Reinforcement Learning for Enhanced Grammatical Error Correction
• CoBA: Counterbias Text Augmentation for Mitigating Various Spurious Correlations via Semantic Triples
• Constraint Matters: Multi-Modal Representation for Reducing Mixed-Integer Linear programming
• Scaling Laws for Task-Stratified Knowledge in Post-Training Quantized Large Language Models
• Principled Detection of Hallucinations in Large Language Models via Multiple Testing
• Test-Time Scaling Strategies for Generative Retrieval in Multimodal Conversational Recommendations
• Exploring Scaling Laws of CTR Model for Online Performance Improvement
• End-to-End On-Device Quantization-Aware Training for LLMs at Inference Cost
• An Empirical Study of Knowledge Distillation for Code Understanding Tasks
• Position Bias Mitigates Position Bias:Mitigate Position Bias Through Inter-Position Knowledge Distillation
• Mapping the Course for Prompt-based Structured Prediction
• Reasoning is about giving reasons
• Vivid-VR: Distilling Concepts from Text-to-Video Diffusion Transformer for Photorealistic Video Restoration
• In2x at WMT25 Translation Task
• Self-Disguise Attack: Induce the LLM to disguise itself for AIGT detection evasion
• MoVieDrive: Multi-Modal Multi-View Urban Scene Video Generation
• Two Birds with One Stone: Multi-Task Detection and Attribution of LLM-Generated Text
• Jointly Extracting Interventions, Outcomes, and Findings from RCT Reports with LLMs
• PENGUIN: Enhancing Transformer with Periodic-Nested Group Attention for Long-term Time Series Forecasting
• SAGA: Learning Signal-Aligned Distributions for Improved Text-to-Image Generation
• A Fully Spectral Neuro-Symbolic Reasoning Architecture with Graph Signal Processing as the Computational Backbone
• Graph Concept Bottleneck Models
• A Language-Signal-Vision Multimodal Framework for Multitask Cardiac Analysis
• Z-Pruner: Post-Training Pruning of Large Language Models for Efficiency without Retraining
• PC-Sampler: Position-Aware Calibration of Decoding Bias in Masked Diffusion Models
• Maximum Score Routing For Mixture-of-Experts
• EgoTwin: Dreaming Body and View in First Person
• RepreGuard: Detecting LLM-Generated Text by Revealing Hidden Representation Patterns
• Improving Detection of Watermarked Language Models
• SEA-BED: Southeast Asia Embedding Benchmark
• MoE-TTS: Enhancing Out-of-Domain Text Understanding for Description-based TTS via Mixture-of-Experts
• Expressive Speech Retrieval using Natural Language Descriptions of Speaking Style
• Copyright Protection for Large Language Models: A Survey of Methods, Challenges, and Trends
• A Dataset for Distilling Knowledge Priors from Literature for Therapeutic Design
• SoK: Data Minimization in Machine Learning
• Neural Machine Translation for Coptic-French: Strategies for Low-Resource Ancient Languages
• FuXi-β: Towards a Lightweight and Fast Large-Scale Generative Recommendation Model
• Semantic IDs for Joint Generative Search and Recommendation
• NanoControl: A Lightweight Framework for Precise and Efficient Control in Diffusion Transformer
• Empowering Multimodal LLMs with External Tools: A Comprehensive Survey
• XQuant: Breaking the Memory Wall for LLM Inference with KV Cache Rematerialization
• Improving Generative Cross-lingual Aspect-Based Sentiment Analysis with Constrained Decoding
• Large Language Models for Summarizing Czech Historical Documents and Beyond
• Advancing Cross-lingual Aspect-Based Sentiment Analysis with LLMs and Constrained Decoding for Sequence-to-Sequence Models
• Meta-Metrics and Best Practices for System-Level Inference Performance Benchmarking
• μ-Parametrization for Mixture of Experts
• GSFixer: Improving 3D Gaussian Splatting with Reference-Guided Video Diffusion Priors
• Improving Dense Passage Retrieval with Multiple Positive Passages
• Utilizing Multilingual Encoders to Improve Large Language Models for Low-Resource Languages
• Lay2Story: Extending Diffusion Transformers for Layout-Togglable Story Generation
• Prompt-Based Approach for Czech Sentiment Analysis
• Leveraging Large Language Models for Rare Disease Named Entity Recognition
• Boosting Action-Information via a Variational Bottleneck on Unlabelled Robot Videos
• Weakly Supervised Fine-grained Span-Level Framework for Chinese Radiology Report Quality Assurance
• Training-Free Text-Guided Color Editing with Multi-Modal Diffusion Transformer
• TARA: Token-Aware LoRA for Composable Personalization in Diffusion Models
• VSF: Simple, Efficient, and Effective Negative Guidance in Few-Step Image Generation Models By Value Sign Flip
• DeCAL Tokenwise Compression
• Cut2Next: Generating Next Shot via In-Context Tuning
• SAEMark: Multi-bit LLM Watermarking with Inference-Time Scaling
• Artificial intelligence and corporate ideation systems
• Towards Multimodal Sentiment Analysis via Contrastive Cross-modal Retrieval Augmentation and Hierachical Prompts
• Retrieval-Augmented Multi-Agent System for Rapid Statement of Work Generation
• Exploring Multimodal Diffusion Transformers for Enhanced Prompt-based Image Editing
• Selective KV-Cache Sharing to Mitigate Timing Side-Channels in LLM Inference
• Large Language Models for Subjective Language Understanding: A Survey
• On Understanding of the Dynamics of Model Capacity in Continual Learning
• GLiClass: Generalist Lightweight Model for Sequence Classification Tasks
• Signature vs. Substance: Evaluating the Balance of Adversarial Resistance and Linguistic Quality in Watermarking Large Language Models
• Czech Dataset for Complex Aspect-Based Sentiment Analysis Tasks
• Enhancing Small-Scale Dataset Expansion with Triplet-Connection-based Sample Re-Weighting
• A Survey on Non-Intrusive ASR Refinement: From Output-Level Correction to Full-Model Distillation
• Improved Personalized Headline Generation via Denoising Fake Interests from Implicit Feedback
• CharacterShot: Controllable and Consistent 4D Character Animation
• Talk2Image: A Multi-Agent System for Multi-Turn Image Generation and Editing
• ESNERA: Empirical and semantic named entity alignment for named entity dataset merging
• Bridging Cultural Nuances in Dialogue Agents through Cultural Value Surveys
• Leveraging LLMs for Smart Cities Qualitative Data Analysis
• Explanatory argument extraction of correct answers in resident medical exams
• What Builds Effective In-Context Examples for Code Generation?
• Bifrost-1: Bridging Multimodal LLMs and Diffusion Models with Patch-level CLIP Latents
• AnomalyMoE: Towards a Language-free Generalist Model for Unified Visual Anomaly Detection
• KnapFormer: An Online Load Balancer for Efficient Diffusion Transformers Training
• DP-LLM: Runtime Model Adaptation with Dynamic Layer-wise Precision Assignment
• Genie Envisioner: A Unified World Foundation Platform for Robotic Manipulation
• PRvL: Quantifying the Capabilities and Risks of Large Language Models for PII Redaction
• Streamlining Admission with LOR Insights: AI-Based Leadership Assessment in Online Master's Program
• Large Language Models Transform Organic Synthesis From Reaction Prediction to Automation
• SONAR-LLM: Autoregressive Transformer that Thinks in Sentence Embeddings and Speaks in Tokens
• Decision-Making with Deliberation: Meta-reviewing as a Document-grounded Dialogue
• A Study of the Framework and Real-World Applications of Language Embedding for 3D Scene Understanding
• Learning from Oblivion: Predicting Knowledge Overflowed Weights via Retrodiction of Forgetting
• REINA: Regularized Entropy Information-Based Loss for Efficient Simultaneous Speech Translation
• iFairy: the First 2-bit Complex LLM with All Parameters in \±1, ± i\
• Fine-Tuning Small Language Models (SLMs) for Autonomous Web-based Geographical Information Systems (AWebGIS)
• Parity-Aware Byte-Pair Encoding: Improving Cross-lingual Fairness in Tokenization
• Live Music Models
• Lightweight Transformers for Zero-Shot and Fine-Tuned Text-to-SQL Generation Using Spider
• HiD-VAE: Interpretable Generative Recommendation via Hierarchical and Disentangled Semantic IDs
• FlexQ: Efficient Post-training INT6 Quantization for LLM Serving via Algorithm-System Co-Design
• Modelling and Classifying the Components of a Literature Review
• Large Language Model's Multi-Capability Alignment in Biomedical Domain
• Gather and Trace: Rethinking Video TextVQA from an Instance-oriented Perspective
• STARE: Predicting Decision Making Based on Spatio-Temporal Eye Movements
• Majority Bit-Aware Watermarking For Large Language Models
• Tackling Distribution Shift in LLM via KILO: Knowledge-Instructed Learning for Continual Adaptation
• Speech-to-LaTeX: New Models and Datasets for Converting Spoken Equations and Sentences
• Draw Your Mind: Personalized Generation via Condition-Level Modeling in Text-to-Image Diffusion Models
• READ: Real-time and Efficient Asynchronous Diffusion for Audio-driven Talking Head Generation
• Cropping outperforms dropout as an augmentation strategy for training self-supervised text embeddings
• Variety Is the Spice of Life: Detecting Misinformation with Dynamic Environmental Representations
• Exploring Layer-wise Information Effectiveness for Post-Training Quantization in Small Language Models
• LLM-Prior: A Framework for Knowledge-Driven Prior Elicitation and Aggregation
• Multimodal Human-Intent Modeling for Contextual Robot-to-Human Handovers of Arbitrary Objects
• LORE: Latent Optimization for Precise Semantic Control in Rectified Flow-based Image Editing
• On the Evaluation of Large Language Models in Multilingual Vulnerability Repair
• Realizing Scaling Laws in Recommender Systems: A Foundation-Expert Paradigm for Hyperscale Model Deployment
• Tricks and Plug-ins for Gradient Boosting with Transformers
• TransAM: Transformer-Based Agent Modeling for Multi-Agent Systems via Local Trajectory Encoding
• LOST: Low-rank and Sparse Pre-training for Large Language Models
• FlashCommunication V2: Bit Splitting and Spike Reserving for Any Bit Communication
• CAMERA: Multi-Matrix Joint Compression for MoE Models via Micro-Expert Redundancy Analysis
• AttriCtrl: Fine-Grained Control of Aesthetic Attribute Intensity in Diffusion Models
• The SMeL Test: A simple benchmark for media literacy in language models
• TRACEALIGN -- Tracing the Drift: Attributing Alignment Failures to Training-Time Belief Sources in LLMs
• Versatile yet Efficient Network Traffic Analysis: Offloading Network Foundation Model to SmartNIC
• A Survey on Data Security in Large Language Models
• MicroMix: Efficient Mixed-Precision Quantization with Microscaling Formats for Large Language Models
• Zero-shot Compositional Action Recognition with Neural Logic Constraints
• SHAMI-MT: A Syrian Arabic Dialect to Modern Standard Arabic Bidirectional Machine Translation System
• MolReasoner: Toward Effective and Interpretable Reasoning for Molecular LLMs
• Quantum-RAG and PunGPT2: Advancing Low-Resource Language Generation and Retrieval for the Punjabi Language
• The Bidirectional Process Reward Model
• EAC-MoE: Expert-Selection Aware Compressor for Mixture-of-Experts Large Language Models
• Empowering Tabular Data Preparation with Language Models: Why and How?
• Am I Blue or Is My Hobby Counting Teardrops? Expression Leakage in Large Language Models as a Symptom of Irrelevancy Disruption
• RouteMark: A Fingerprint for Intellectual Property Attribution in Routing-based Model Merging
• Training Dynamics of the Cooldown Stage in Warmup-Stable-Decay Learning Rate Scheduler
• TreeDiff: AST-Guided Code Generation with Diffusion LLMs
• StyDeco: Unsupervised Style Transfer with Distilling Priors and Semantic Decoupling
• A Note on Code Quality Score: LLMs for Maintainable Large Codebases
• AudioGen-Omni: A Unified Multimodal Diffusion Transformer for Video-Synchronized Audio, Speech, and Song Generation
• From Generator to Embedder: Harnessing Innate Abilities of Multimodal LLMs via Building Zero-Shot Discriminative Embedding Model
• LAMIC: Layout-Aware Multi-Image Composition via Scalability of Multimodal Diffusion Transformer
• DACTYL: Diverse Adversarial Corpus of Texts Yielded from Large Language Models
• ART: Adaptive Relation Tuning for Generalized Relation Prediction
• Causal2Vec: Improving Decoder-only LLMs as Versatile Embedding Models
• Text-to-SQL Task-oriented Dialogue Ontology Construction
• What's Taboo for You? - An Empirical Evaluation of LLMs Behavior Toward Sensitive Content
• SequenceLayers: Sequence Processing and Streaming Neural Networks Made Easy
• Unveiling Super Experts in Mixture-of-Experts Large Language Models
• How Far Are AI Scientists from Changing the World?
• From Image Captioning to Visual Storytelling
• Multi-Modal Motion Retrieval by Learning a Fine-Grained Joint Embedding Space
• H-RDT: Human Manipulation Enhanced Bimanual Robotic Manipulation
• Segment Anything for Video: A Comprehensive Review of Video Object Segmentation and Tracking from Past to Future
• A Systematic Literature Review on Detecting Software Vulnerabilities with Large Language Models
• XAutoLM: Efficient Fine-Tuning of Language Models via Meta-Learning and AutoML
• On the Reliability of Vision-Language Models Under Adversarial Frequency-Domain Perturbations
• SAEL: Leveraging Large Language Models with Adaptive Mixture-of-Experts for Smart Contract Vulnerability Detection
• A Comprehensive Taxonomy of Negation for NLP and Neural Retrievers
• Uni-Mol3: A Multi-Molecular Foundation Model for Advancing Organic Reaction Modeling
• Context-aware Rotary Position Embedding
• Generative Recommendation with Semantic IDs: A Practitioner's Handbook
• Fine-Tuning Code Language Models to Detect Cross-Language Bugs
• Towards Locally Deployable Fine-Tuned Causal Large Language Models for Mode Choice Behaviour
• DiTalker: A Unified DiT-based Framework for High-Quality and Speaking Styles Controllable Portrait Animation
• Turbocharging Web Automation: The Impact of Compressed History States
• Bubbleformer: Forecasting Boiling with Transformers
• Memorization in Fine-Tuned Large Language Models

Discussions

• Exploring the Limits of Transfer Learning with a Unified Transformer (2019) [hn, 12 points, 1 comments]
• another classic from the "this sounds too stupid to actually work, but we swear it does" shelf is the T5 paper arxiv.org/abs/1910.10683 [bsky, 1 points, 0 comments]
• T5: Encoder-only or decoder-only is NOT all you need, though text-to-text is all you need. (Also, pre-training + finetuning 🚀) https://arxiv.org/abs/1910.10683 [bsky, 0 points, 1 comments]
• 🎧 EP012: Google T5 Turns Every Task Into Text 📄 T5 🔗 https://arxiv.org/abs/1910.10683 🟢 https://podcasters.spotify.com/pod/show/yun-wu/episodes/EP012-Google-T5-Turns-Every-Task-Into-Text-e3fin1c ▶ [bsky, 0 points, 0 comments]

Related

• A Survey on Transfer Learning in Natural Language Processing
• Evolution of transfer learning in natural language processing
• Towards a Unified View of Parameter-Efficient Transfer Learning
• Exploring and Predicting Transferability across NLP Tasks
• Using Transfer Learning for Code-Related Tasks