NeurIPS 2026 reviews are up: reviews, scores, rebuttals and meta-reviews for 7,877 papers. See the highest rated
acceptodds: the papers of ICLR 2027
119 results for “keyword:"vision transformers"” in ICLR 2027. 43 withdrawn or desk rejected listed apart. clearsort:relevanceratingnewestmost read
[PDF]Reweighting Framewise Attention in Video Transformers for Facial Expression Understanding▲3952.67
[PDF]Do Vision Transformers Need All-to-All Attention? Global Communication Through Elastic Learned Cores▲143.33
[PDF]The Representation-Continuation Gap: Low-Rank States vs. Low-Rank Dynamics in Vision Transformers▼1.2k1.67
[PDF]From Sequential Refinement to One-Shot Global Selection: Monotone Adaptive Tokenization▼4.5k2.67
Withdrawn or desk rejected (43)
[PDF]Cell-level representations in pathology foundation models are a question of read-outWithdrawn1.50
[PDF]BitRot: Understanding Quantization Fragility in Vision Transformers Under Distribution ShiftWithdrawn2.00