{"total":77,"rows":[{"id":"2504.13181","title":"Perception Encoder: The best visual embeddings are not at the output of the network","created":"2025-04-17","cited_by_count":326,"area":null},{"id":"2410.07073","title":"Pixtral 12B","created":"2024-10-09","cited_by_count":179,"area":"mllm"},{"id":"2310.15308","title":"SAM-CLIP: Merging Vision Foundation Models towards Semantic and Spatial Understanding","created":"2023-10-23","cited_by_count":111,"area":null},{"id":"2307.13721","title":"Foundational Models Defining a New Era in Vision: A Survey and Outlook","created":"2023-07-25","cited_by_count":110,"area":null},{"id":"2312.08578","title":"A Picture is Worth More Than 77 Text Tokens: Evaluating CLIP-Style Models on Dense Captions","created":"2023-12-14","cited_by_count":108,"area":null},{"id":"2311.17049","title":"MobileCLIP: Fast Image-Text Models through Multi-Modal Reinforced Training","created":"2023-11-28","cited_by_count":96,"area":null},{"id":"2406.08478","title":"What If We Recaption Billions of Web Images with LLaMA-3?","created":"2024-06-12","cited_by_count":87,"area":null},{"id":"2509.08826","title":"RewardDance: Reward Scaling in Visual Generation","created":"2025-09-10","cited_by_count":68,"area":"alignment"},{"id":"2404.07983","title":"Two Effects, One Trigger: On the Modality Gap, Object Bias, and Information Imbalance in Contrastive Vision-Language Models","created":"2024-04-11","cited_by_count":57,"area":"mllm"},{"id":"2505.18657","title":"MLLMs are Deeply Affected by Modality Bias","created":"2025-05-24","cited_by_count":40,"area":"mllm"},{"id":"2404.02132","title":"ViTamin: Designing Scalable Vision Models in the Vision-Language Era","created":"2024-04-02","cited_by_count":39,"area":"mllm"},{"id":"2407.19584","title":"SaulLM-54B & SaulLM-141B: Scaling Up Domain Adaptation for the Legal Domain","created":"2024-07-28","cited_by_count":39,"area":null},{"id":"2410.00263","title":"Procedure-Aware Surgical Video-language Pretraining with Hierarchical Knowledge Augmentation","created":"2024-09-30","cited_by_count":39,"area":null},{"id":"2401.04578","title":"Effective pruning of web-scale datasets based on complexity of concept clusters","created":"2024-01-09","cited_by_count":34,"area":null},{"id":"2401.04727","title":"Revisiting Adversarial Training at Scale","created":"2024-01-09","cited_by_count":33,"area":null},{"id":"2406.17711","title":"Data curation via joint example selection further accelerates multimodal learning","created":"2024-06-25","cited_by_count":33,"area":null},{"id":"2408.04810","title":"UniBench: Visual Reasoning Requires Rethinking Vision-Language Beyond Scaling","created":"2024-08-09","cited_by_count":31,"area":"lr"},{"id":"2504.10462","title":"The Scalability of Simplicity: Empirical Analysis of Vision-Language Learning with a Single Transformer","created":"2025-04-14","cited_by_count":30,"area":"mllm"},{"id":"2405.13226","title":"Dataset Decomposition: Faster LLM Training with Variable Sequence Length Curriculum","created":"2024-05-21","cited_by_count":26,"area":null},{"id":"2306.15658","title":"CLIPA-v2: Scaling CLIP Training with 81.1% Zero-shot ImageNet Accuracy within a \\$10,000 Budget; An Extra \\$4,000 Unlocks 81.8% Accuracy","created":"2023-06-27","cited_by_count":24,"area":null},{"id":"2404.16030","title":"MoDE: CLIP Data Experts via Clustering","created":"2024-04-24","cited_by_count":22,"area":null},{"id":"2505.05528","title":"X-Transfer Attacks: Towards Super Transferable Adversarial Attacks on CLIP","created":"2025-05-08","cited_by_count":21,"area":"mllm"},{"id":"2508.20691","title":"MobileCLIP2: Improving Multi-Modal Reinforced Training","created":"2025-08-28","cited_by_count":21,"area":"distill"},{"id":"2402.07417","title":"An Empirical Study Into What Matters for Calibrating Vision-Language Models","created":"2024-02-12","cited_by_count":20,"area":"mllm"},{"id":"2312.02147","title":"Rejuvenating image-GPT as Strong Visual Representation Learners","created":"2023-12-04","cited_by_count":16,"area":null},{"id":"2505.04601","title":"OpenVision: A Fully-Open, Cost-Effective Family of Advanced Vision Encoders for Multimodal Learning","created":"2025-05-07","cited_by_count":16,"area":null},{"id":"2402.09816","title":"Mind the Modality Gap: Towards a Remote Sensing Vision-Language Model via Cross-modal Alignment","created":"2024-02-15","cited_by_count":15,"area":"mllm"},{"id":"2403.16442","title":"If CLIP Could Talk: Understanding Vision-Language Model Representations Through Their Preferred Concept Descriptions","created":"2024-03-25","cited_by_count":14,"area":"mllm"},{"id":"2410.11087","title":"Locality Alignment Improves Vision-Language Models","created":"2024-10-14","cited_by_count":13,"area":"lr"},{"id":"2502.18176","title":"CLIPure: Purification in Latent Space via CLIP for Adversarially Robust Zero-Shot Classification","created":"2025-02-25","cited_by_count":13,"area":null},{"id":"2504.04893","title":"SCAM: A Real-World Typographic Robustness Evaluation for Multimodal Foundation Models","created":"2025-04-07","cited_by_count":13,"area":"mllm"},{"id":"2411.11927","title":"FLAME: Frozen Large Language Models Enable Data-Efficient Language-Image Pre-training","created":"2024-11-18","cited_by_count":12,"area":null},{"id":"2311.17833","title":"DiG-IN: Diffusion Guidance for Investigating Networks -- Uncovering Classifier Differences Neuron Visualisations and Visual Counterfactual Explanations","created":"2023-11-29","cited_by_count":11,"area":null},{"id":"2411.03313","title":"Classification Done Right for Vision-Language Pre-Training","created":"2024-11-05","cited_by_count":10,"area":"mllm"},{"id":"2501.09446","title":"Double Visual Defense: Adversarial Pre-training and Instruction Tuning for Improving Vision-Language Model Robustness","created":"2025-01-16","cited_by_count":10,"area":"mllm"},{"id":"2401.07770","title":"Seeing the Unseen: Visual Common Sense for Semantic Placement","created":"2024-01-15","cited_by_count":9,"area":null},{"id":"2405.20299","title":"Scaling White-Box Transformers for Vision","created":"2024-05-30","cited_by_count":9,"area":null},{"id":"2407.01445","title":"FastCLIP: A Suite of Optimization Techniques to Accelerate CLIP Training with Limited Resources","created":"2024-07-01","cited_by_count":9,"area":null},{"id":"2410.10308","title":"LG-CAV: Train Any Concept Activation Vector with Language Guidance","created":"2024-10-14","cited_by_count":9,"area":"mllm"},{"id":"2411.11667","title":"Dissecting Representation Misalignment in Contrastive Learning via Influence Function","created":"2024-11-18","cited_by_count":8,"area":null},{"id":"2401.00028","title":"An Empirical Study of Scaling Law for OCR","created":"2023-12-29","cited_by_count":7,"area":null},{"id":"2502.09598","title":"GAIA: A Global, Multi-modal, Multi-scale Vision-Language Dataset for Remote Sensing Image Analysis","created":"2025-02-13","cited_by_count":7,"area":"mllm"},{"id":"2502.15809","title":"Black Sheep in the Herd: Playing with Spuriously Correlated Attributes for Vision-Language Recognition","created":"2025-02-19","cited_by_count":7,"area":"mllm"},{"id":"2503.17080","title":"Seeing What Matters: Empowering CLIP with Patch Generation-to-Selection","created":"2025-03-21","cited_by_count":7,"area":null},{"id":"2509.01644","title":"OpenVision 2: A Family of Generative Pretrained Visual Encoders for Multimodal Learning","created":"2025-09-01","cited_by_count":7,"area":null},{"id":"2410.13694","title":"Exploring the Design Space of Visual Context Representation in Video MLLMs","created":"2024-10-17","cited_by_count":6,"area":"mllm"},{"id":"2411.09126","title":"SCAN: Bootstrapping Contrastive Pre-training for Data Efficiency","created":"2024-11-14","cited_by_count":6,"area":null},{"id":"2505.24424","title":"Advancing Compositional Awareness in CLIP with Efficient Fine-Tuning","created":"2025-05-30","cited_by_count":5,"area":"mllm"},{"id":"2601.15369","title":"OpenVision 3: A Family of Unified Visual Encoder for Both Understanding and Generation","created":"2026-01-21","cited_by_count":5,"area":null},{"id":"2508.04211","title":"What Holds Back Open-Vocabulary Segmentation?","created":"2025-08-06","cited_by_count":4,"area":null},{"id":"2311.18082","title":"Zooming Out on Zooming In: Advancing Super-Resolution for Remote Sensing","created":"2023-11-29","cited_by_count":3,"area":null},{"id":"2401.11633","title":"Zoom-shot: Fast and Efficient Unsupervised Zero-Shot Transfer of CLIP to Vision Encoders with Multimodal Loss","created":"2024-01-22","cited_by_count":3,"area":"mllm"},{"id":"2501.04784","title":"Leveraging Registers in Vision Transformers for Robust Adaptation","created":"2025-01-08","cited_by_count":3,"area":null},{"id":"2502.03950","title":"LR0.FM: Low-Res Benchmark and Improving Robustness for Zero-Shot Classification in Foundation Models","created":"2025-02-06","cited_by_count":3,"area":null},{"id":"2511.08417","title":"NeuCLIP: Efficient Large-Scale CLIP Training with Neural Normalizer Optimization","created":"2025-11-11","cited_by_count":3,"area":null},{"id":"2404.00509","title":"DailyMAE: Towards Pretraining Masked Autoencoders in One Day","created":"2024-03-31","cited_by_count":2,"area":null},{"id":"2410.11963","title":"CtrlSynth: Controllable Image Text Synthesis for Data-Efficient Multimodal Learning","created":"2024-10-15","cited_by_count":2,"area":null},{"id":"2505.18983","title":"AmorLIP: Efficient Language-Image Pretraining via Amortization","created":"2025-05-25","cited_by_count":2,"area":null},{"id":"2508.20760","title":"Occlusion Robustness of CLIP for Military Vehicle Classification","created":"2025-08-28","cited_by_count":2,"area":"mllm"},{"id":"2403.18525","title":"Language Plays a Pivotal Role in the Object-Attribute Compositional Generalization of CLIP","created":"2024-03-27","cited_by_count":1,"area":"mllm"},{"id":"2407.15487","title":"In-Context Learning Improves Compositional Understanding of Vision-Language Models","created":"2024-07-22","cited_by_count":1,"area":"mllm"},{"id":"2409.13513","title":"Efficient and Discriminative Image Feature Extraction for Universal Image Retrieval","created":"2024-09-20","cited_by_count":1,"area":null},{"id":"2412.16148","title":"Frequency Is What You Need: Considering Word Frequency When Text Masking Benefits Vision-Language Model Pre-training","created":"2024-12-20","cited_by_count":1,"area":"mllm"},{"id":"2509.02175","title":"Understanding Space Is Rocket Science -- Only Top Reasoning Models Can Solve Spatial Understanding Tasks","created":"2025-09-02","cited_by_count":1,"area":"lr"},{"id":"2512.11982","title":"Semantic search for 100M+ galaxy images using AI-generated captions","created":"2025-12-12","cited_by_count":1,"area":null},{"id":"2412.03871","title":"CLIP-PING: Boosting Lightweight Vision-Language Models with Proximus Intrinsic Neighbors Guidance","created":"2024-12-05","cited_by_count":0,"area":"mllm"},{"id":"2507.03300","title":"LRM-1B: Towards Large Routing Model","created":"2025-07-04","cited_by_count":0,"area":null},{"id":"2510.19559","title":"A Matter of Time: Revealing the Structure of Time in Vision-Language Models","created":"2025-10-22","cited_by_count":0,"area":"mllm"},{"id":"2603.25722","title":"No Hard Negatives Required: Concept Centric Learning Leads to Compositionality without Degrading Zero-shot Capabilities of Contrastive Models","created":"2026-03-26","cited_by_count":0,"area":null},{"id":"2604.10064","title":"On The Application of Linear Attention in Multimodal Transformers","created":"2026-04-11","cited_by_count":0,"area":"mllm"},{"id":"2604.27932","title":"Dynamic Cluster Data Sampling for Efficient and Long-Tail-Aware Vision-Language Pre-training","created":"2026-04-30","cited_by_count":0,"area":"mllm"},{"id":"2605.18029","title":"What Matters for Grocery Product Retrieval with Open Source Vision Language Models","created":"2026-05-18","cited_by_count":0,"area":"mllm"},{"id":"2606.26617","title":"Sketched Linear Contrastive Learning: Approximation, Optimization, and Statistical Scaling","created":"2026-06-25","cited_by_count":0,"area":null},{"id":"2607.28211","title":"Scaling Vision-Language Models Is Not Enough to Mitigate Bias","created":"2026-07-30","cited_by_count":0,"area":"mllm"},{"id":"2609.05730","title":"Bigger Text Encoders Can Hurt CLIP Zero-Shot Performance","created":"2026-09-04","cited_by_count":0,"area":null},{"id":"2609.35583","title":"What Paired Evaluations Reveal under Visual Perturbations","created":"2026-09-28","cited_by_count":0,"area":null},{"id":"2610.00447","title":"Frozen Scenes, Shifting Winners: Configuration Fragility in Text-to-3D Evaluation","created":"2026-09-30","cited_by_count":0,"area":null}],"next_cursor":null}