English
Related papers

Related papers: Not All Attention Heads Are What You Need: Refinin…

200 papers

We investigate the CLIP image encoder by analyzing how individual model components affect the final representation. We decompose the image representation as a sum across individual image patches, model layers, and attention heads, and use…

Computer Vision and Pattern Recognition · Computer Science 2024-04-01 Yossi Gandelsman , Alexei A. Efros , Jacob Steinhardt

Recent neural models for image captioning usually employ an encoder-decoder framework with an attention mechanism. However, the attention mechanism in such a framework aligns one single (attended) image feature vector to one caption word,…

Computer Vision and Pattern Recognition · Computer Science 2020-01-07 Lun Huang , Wenmin Wang , Yaxian Xia , Jie Chen

Multimodal models like CLIP have gained significant attention due to their remarkable zero-shot performance across various tasks. However, studies have revealed that CLIP can inadvertently learn spurious associations between target…

Computer Vision and Pattern Recognition · Computer Science 2025-05-26 Wei Jie Yeo , Rui Mao , Moloud Abdar , Erik Cambria , Ranjan Satapathy

Standard fairness audits of foundation models quantify that a model is biased, but not where inside the network the bias resides. We propose a mechanistic fairness audit that combines projected residual-stream decomposition, zero-shot…

Computer Vision and Pattern Recognition · Computer Science 2026-03-13 Alaa Yasser , Kittipat Phunjanna , Marcos Escudero Viñolo , Catarina Barata , Jenny Benois-Pineau

Mechanistic interpretability improves the safety, reliability, and robustness of large AI models. This study examined individual attention heads in vision transformers (ViTs) fine tuned on distorted 2D spectrogram images containing non…

Machine Learning · Computer Science 2025-03-25 Nooshin Bahador

Domain Generalization (DG) aims to learn a model from multiple source domains to achieve satisfactory performance on unseen target domains. Recent works introduce CLIP to DG tasks due to its superior image-text alignment and zeros-shot…

Computer Vision and Pattern Recognition · Computer Science 2024-12-11 Yingfan Wang , Guoliang Kang

Recent work has explored how individual components of the CLIP-ViT model contribute to the final representation by leveraging the shared image-text representation space of CLIP. These components, such as attention heads and MLPs, have been…

Computer Vision and Pattern Recognition · Computer Science 2024-10-22 Sriram Balasubramanian , Samyadeep Basu , Soheil Feizi

Vision transformer (ViT) expands the success of transformer models from sequential data to images. The model decomposes an image into many smaller patches and arranges them into a sequence. Multi-head self-attentions are then applied to the…

Machine Learning · Computer Science 2023-03-27 Yiran Li , Junpeng Wang , Xin Dai , Liang Wang , Chin-Chia Michael Yeh , Yan Zheng , Wei Zhang , Kwan-Liu Ma

We present a novel technique for interpreting the neurons in CLIP-ResNet by decomposing their contributions to the output into individual computation paths. More specifically, we analyze all pairwise combinations of neurons and the…

Computer Vision and Pattern Recognition · Computer Science 2025-12-02 Edmund Bu , Yossi Gandelsman

Extending CLIP models to semantic segmentation remains challenging due to the misalignment between their image-level pre-training objectives and the pixel-level visual understanding required for dense prediction. While prior efforts have…

Computer Vision and Pattern Recognition · Computer Science 2025-10-29 Jinxin Zhou , Jiachen Jiang , Zhihui Zhu

Recent Vision-Language Models (VLMs) have demonstrated remarkable multimodal understanding capabilities, yet the redundant visual tokens incur prohibitive computational overhead and degrade inference efficiency. Prior studies typically…

Computer Vision and Pattern Recognition · Computer Science 2026-02-24 Qiankun Ma , Ziyao Zhang , Haofei Wang , Jie Chen , Zhen Song , Hairong Zheng

CLIP is one of the most popular foundational models and is heavily used for many vision-language tasks. However, little is known about the inner workings of CLIP. To bridge this gap we propose a study to quantify the interpretability in…

Computer Vision and Pattern Recognition · Computer Science 2024-09-11 Avinash Madasu , Yossi Gandelsman , Vasudev Lal , Phillip Howard

Contrastive Language-Image Pre-training (CLIP) plays an essential role in extracting valuable content information from images across diverse tasks. It aligns textual and visual modalities to comprehend the entire image, including all the…

Computer Vision and Pattern Recognition · Computer Science 2023-12-15 Zeyi Sun , Ye Fang , Tong Wu , Pan Zhang , Yuhang Zang , Shu Kong , Yuanjun Xiong , Dahua Lin , Jiaqi Wang

CLIP is one of the most popular foundation models and is heavily used for many vision-language tasks, yet little is known about its inner workings. As CLIP is increasingly deployed in real-world applications, it is becoming even more…

Computer Vision and Pattern Recognition · Computer Science 2025-09-22 Avinash Madasu , Vasudev Lal , Phillip Howard

Contrastive vision-language models excel in zero-shot image recognition but face challenges in few-shot scenarios due to computationally intensive offline fine-tuning using prompt learning, which risks overfitting. To overcome these…

Computer Vision and Pattern Recognition · Computer Science 2025-09-05 Phuoc-Nguyen Bui , Khanh-Binh Nguyen , Hyunseung Choo

Zero-shot anomaly detection (ZSAD) enables anomaly detection without normal samples from target categories, addressing scenarios where task-specific training data is unavailable. However, existing ZSAD methods either neglect adaptation of…

Computer Vision and Pattern Recognition · Computer Science 2026-02-03 Kiyoon Jeong , Jaehyuk Heo , Junyeong Son , Pilsung Kang

Despite the success of large-scale pretrained Vision-Language Models (VLMs) especially CLIP in various open-vocabulary tasks, their application to semantic segmentation remains challenging, producing noisy segmentation maps with…

Computer Vision and Pattern Recognition · Computer Science 2024-07-18 Mengcheng Lan , Chaofeng Chen , Yiping Ke , Xinjiang Wang , Litong Feng , Wayne Zhang

Large Vision-Language Models (VLMs) have achieved remarkable success in multi-modal reasoning, but their inference time efficiency remains a significant challenge due to the memory overhead during decoding, especially when the query and…

Computer Vision and Pattern Recognition · Computer Science 2026-03-26 Fatih Ilhan , Gaowen Liu , Ramana Rao Kompella , Selim Furkan Tekin , Tiansheng Huang , Zachary Yahn , Yichang Xu , Ling Liu

The popularity of Contrastive Language-Image Pre-training (CLIP) has propelled its application to diverse downstream vision tasks. To improve its capacity on downstream tasks, few-shot learning has become a widely-adopted technique.…

Computer Vision and Pattern Recognition · Computer Science 2023-04-04 Xiangyang Zhu , Renrui Zhang , Bowei He , Aojun Zhou , Dong Wang , Bin Zhao , Peng Gao

Understanding the representation shift on Vision Language Models like CLIP under different augmentations provides valuable insights on Mechanistic Interpretability. In this study, we show the shift on CLIP's embeddings on 9 common…

Computer Vision and Pattern Recognition · Computer Science 2026-03-24 Ashim Dahal , Saydul Akbar Murad , Nick Rahimi
‹ Prev 1 2 3 10 Next ›