No article text was recorded for this version.
compositional reasoning failure (compositional reasoning gap)
Version 1 of 184 papers95 conceptsfrozen 2026-10-04
Papers in this version
- Blog: Nouns, Not Modifiers: OpenVLA Parses Objects but Fails at Spatial Composition
- Blog: PUG: Photorealistic and Semantically Controllable Synthetic Data for Representation Learning
- Blog: Text encoders bottleneck compositionality in contrastive vision-language models
- Blog: Can VLMs Actually See and Read? A Survey on Modality Collapse in Vision-Language Models
- Blog: Measuring Progress in Fine-grained Vision-and-Language Understanding
- Blog: TokenSwap: Backdoor Attack on the Compositional Understanding of Large Vision-Language Models
- Blog: Multilingual Vision-Language Pre-training for the Remote Sensing Domain
- Blog: Do Vision-Language Models Really Understand Visual Language?
- Blog: MetaCLUE: Towards Comprehensive Visual Metaphors Research
- Blog: Towards Unsupervised Image Captioning With Shared Multimodal Embeddings
- Blog: Building Vision-Language Models on Solid Foundations with Masked Distillation
- Blog: Test-Time Matching: Unlocking Compositional Reasoning in Multimodal Models
- Blog: An Inverse Scaling Law for CLIP Training
- Blog: What You See is What You Read? Improving Text-Image Alignment Evaluation
- Blog: Preserving Multi-Modal Capabilities of Pre-trained VLMs for Improving Vision-Linguistic Compositionality
- Blog: Adaptive Testing of Computer Vision Models
- Blog: Enhancing Vision-Language Compositional Understanding with Multimodal Synthetic Data
- Blog: CompoDistill: Attention Distillation for Compositional Reasoning in Multimodal LLMs
- Blog: Open3DSG: Open-Vocabulary 3D Scene Graphs from Point Clouds with Queryable Objects and Open-Set Relationships
- Blog: Global Geometry Is Not Enough for Vision Representations
- Blog: Zero-shot Referring Expression Comprehension via Structural Similarity Between Images and Captions
- Blog: 3VL: using Trees to teach Vision & Language models compositional concepts
- Blog: Teaching CLIP to Count to Ten
- Blog: Fine-tuning CLIP Text Encoders with Two-step Paraphrasing
- Blog: Discffusion: Discriminative Diffusion Models as Few-shot Vision and Language Learners
- Blog: ICSVR: Investigating Compositional and Syntactic Understanding in Video Retrieval Models
- Blog: ImageInWords: Unlocking Hyper-Detailed Image Descriptions
- Blog: Winoground: Probing Vision and Language Models for Visio-Linguistic Compositionality
- Blog: On Erroneous Agreements of CLIP Image Embeddings
- Blog: iSEARLE: Improving Textual Inversion for Zero-Shot Composed Image Retrieval
- Blog: Your Diffusion Model is Secretly a Zero-Shot Classifier
- Blog: Iterated Learning Improves Compositionality in Large Vision-Language Models
- Blog: The Hard Positive Truth About Vision-Language Compositionality
- Blog: Diffusion Classifiers Understand Compositionality, but Conditions Apply
- Blog: In-Context Learning Improves Compositional Understanding of Vision-Language Models
- Blog: Kosmos-G: Generating Images in Context with Multimodal Large Language Models
- Blog: CoVLM: Composing Visual Entities and Relationships in Large Language Models Via Communicative Decoding
- Blog: SPARO: Selective Attention for Robust and Compositional Transformer Encodings for Vision
- Blog: When does perceptual alignment benefit vision representations?
- Blog: COMPACT: COMPositional Atomic-to-Complex Visual Capability Tuning
- Blog: Test-Time Matching: Unlocking Compositional Reasoning in Multimodal Models
- Blog: When and Why Vision-Language Models Behave like Bags-Of-Words, and What to Do About It?
- Blog: Exploring the Potential of Dense Information in Multimodal Alignment
- Blog: Exploring How Generative MLLMs Perceive More Than CLIP with the Same Vision Encoder
- Blog: Text encoders are performance bottlenecks in contrastive vision-language models
- Blog: Incorporating Structured Representations into Pretrained Vision & Language Models Using Scene Graphs
- Blog: Scalable Performance Analysis for Vision-Language Models
- Blog: A Corpus for Reasoning about Natural Language Grounded in Photographs
- Blog: lilGym: Natural Language Visual Reasoning with Reinforcement Learning
- Blog: ConStruct-VL: Data-Free Continual Structured VL Concepts Learning
- Blog: Understanding the Effect of using Semantically Meaningful Tokens for Visual Representation Learning
- Blog: New Dataset and Methods for Fine-Grained Compositional Referring Expression Comprehension via Specialist-MLLM Collaboration
- Blog: Fine-tuning CLIP Text Encoders with Two-step Paraphrasing
- Blog: Your Diffusion Model is Secretly a Zero-Shot Classifier
- Blog: TULIP: Towards Unified Language-Image Pretraining
- Blog: Compositional Chain-of-Thought Prompting for Large Multimodal Models
- Blog: Annotations Are Not All You Need: A Cross-modal Knowledge Transfer Network for Unsupervised Temporal Sentence Grounding
- Blog: Removing Distributional Discrepancies in Captions Improves Image-Text Alignment
- Blog: Contrasting Intra-Modal and Ranking Cross-Modal Hard Negatives to Enhance Visio-Linguistic Compositional Understanding
- Blog: LERF: Language Embedded Radiance Fields
Showing 60 of 84 papers.
History & citation
Current · Permanent link
- This direction split from Compositional Reasoning in Vision Language Models (90% overlap)
Cite this direction
Latest version (a living page):
@misc{lacuna_compositional_reasoning_failure_composit_a939afaa,
title = {compositional reasoning failure (compositional reasoning gap)},
author = {{Lacuna}},
howpublished = {Lacuna research direction},
year = {2026},
url = {https://lacuna.tiptreesystems.com/direction/compositional-reasoning-failure-compositional-reasoning-gap/rdr_a939afaac0d34e2fb75bd5733bdbcda9},
note = {Latest version of a living research direction; cite a version for a fixed text. Accessed 2026-11-22},
}This exact version (version 1, frozen text):
@misc{lacuna_compositional_reasoning_failure_composit_a939afaa_v1,
title = {compositional reasoning failure (compositional reasoning gap)},
author = {{Lacuna}},
howpublished = {Lacuna research direction},
year = {2026},
url = {https://lacuna.tiptreesystems.com/direction/compositional-reasoning-failure-compositional-reasoning-gap/rdr_a939afaac0d34e2fb75bd5733bdbcda9/v/1},
note = {Version 1, frozen 2026-10-04. Accessed 2026-11-22},
}