AI RESEARCH
Vision Transformers Need More Than Registers
arXiv CS.CV
•
ArXi:2602.22394v2 Announce Type: replace Vision Transformers (ViTs), when pre-trained on large-scale data, provide general-purpose representations for diverse downstream tasks. However, artifacts in ViTs are widely observed across different supervision paradigms and downstream tasks. Through systematic analysis of artifacts in ViTs, we find that their fundamental mechanisms have yet to be sufficiently elucidated.