@article{dyllmefficientdiffusionllminferenceviasa,
  title = {DyLLM: Efficient Diffusion LLM Inference via Saliency-based Token Selection and Partial Attention},
  author = {Younjoo Lee and Seungkyun Dan and Junghoo Lee and Jaiyoung Park and Jung Ho Ahn},
  year = {2026},
  eprint = {2603.08026},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2603.08026},
}