The field of deep learning is witnessing significant advancements in model merging and knowledge transfer. Researchers are exploring innovative methods to integrate knowledge from multiple models, enabling more robust and flexible models. One notable direction is the development of frameworks that model the input-representation space, allowing for more effective merging of task-specific knowledge. Additionally, there is a growing interest in leveraging large model repositories to facilitate knowledge transfer between models, leading to improved performance on various tasks. Noteworthy papers in this area include: Model Merging with Functional Dual Anchors, which proposes a novel framework for model merging, and Simplifying Knowledge Transfer in Pretrained Models, which introduces a data partitioning strategy for autonomous knowledge transfer between models. Alias-Free ViT and Implicit Modeling for Transferability Estimation of Vision Foundation Models also demonstrate significant contributions to the field, with the former proposing a shift-invariant Vision Transformer and the latter introducing a novel framework for transferability estimation.