arXiv:2607. 18363v1 Announce Type: cross Abstract: Feed-forward networks hold two thirds of a transformer's non-embedding parameters, yet the architecture has not received a necessity test that controls parameters, compute, and depth at once.
arXiv:2606. 16429v1 Announce Type: new Abstract: Hybrid linear attention models offer an appealing path to faster long-context inference: they reduce the quadratic cost and KV-cache burden of full softmax attention while retaining much of the quality of Transformer models.
arXiv:2602. 06883v3 Announce Type: replace Abstract: The smoothness of the transformer architecture has been extensively studied in the context of generalization, training stability, and adversarial robustness.
The paper introduces CODI, a method that enhances self‑supervised Vision Transformers by mixing a small portion of the teacher’s class‑token distribution into each teacher patch‑token target. This cross distillation approach gives every masked patch an attention‑independent semantic correction without altering the architecture. Experiments demonstrate that CODI improves performance on both dense prediction and image‑level tasks under aligned pretraining settings.
Self-supervised Vision Transformers learn strong image-level representations, but their patch-level representations can be unreliable for dense prediction. We study this mismatch and find that the cla...
arXiv:2607.09086v2 Announce Type: replace Abstract: We present Subtoken Vision Transformer (SubViT), a selective image tokenization method for fine-grained visual recognition. Standard Vision Transfo...
The projection of queries and keys are central to the attention mechanism in Transformer architectures. While they are mathematically symmetric, they play different roles in attention mechanisms.
The paper introduces FAD‑ViT, a vision transformer that decouples spatial and feature axes of global attention, making feature‑axis attention the primary operator. Experiments on 23 medical imaging tasks show that shifting global attention from spatial to feature leaves from‑scratch performance unchanged but significantly boosts transfer learning, achieving up to three times the gain with fewer parameters and lower compute. The study demonstrates that prioritizing feature‑axis attention is a simple, cost‑effective design choice for improving medical image transfer.
arXiv:2606. 14757v1 Announce Type: cross Abstract: Though Vision Transformers (ViTs) have become the dominant backbone in many computer vision tasks, due to permutation equivariance, their attention mechanism lacks explicit spatial inductive biases.
arXiv:2607. 00434v1 Announce Type: cross Abstract: Vision-language models (VLMs) have become a paradigm for multimodal learning, yet remain unstable due to object hallucination, weak visual grounding, and catastrophic forgetting after full-parameter instruction tuning.
arXiv:2609.36448v1 Announce Type: new Abstract: Transformers have demonstrated remarkable in-context learning (ICL) capabilities, enabling them to perform new tasks without additional fine-tuning. Ho...
arXiv:2610.15855v1 Announce Type: new Abstract: We study how to detect and mitigate visual attention sinks in vision-language models (VLMs). Visual attention sinks are image tokens that absorb a larg...