Songlin Yang

profile.jpg

Songlin (松琳) is a second-year PhD student at MIT CSAIL, advised by Prof. Yoon Kim. She earned her bachelor’s degree from SUSTech in 2020 and her master’s degree from ShanghaiTech in 2023, where she was advised by Prof. Kewei Tu.

Her research focuses on the intersection of machine learning systems and large language models, with a particular interest in hardware-aware algorithm design for efficient sequence modeling — especially in linear attention models. For more on her work, see this video and her slides.

She is a strong advocate for open-source research 🐳. Explore the open-source library flash-linear-attention and the fully virtual seminar series Advances in Sequence Modeling from Algorithmic Perspectives — past talks are available here. The best way to reach her is through the FLA Discord Community.

latest posts

selected publications

  1. ICLR
    Gated Delta Networks: Improving Mamba2 with Delta Rule
    Songlin Yang, Jan Kautz, and Ali Hatamizadeh
    2025
  2. NeurIPS
    Parallelizing Linear Transformers with the Delta Rule over Sequence Length
    Songlin Yang, Bailin WangYu ZhangYikang Shen, and Yoon Kim
    In , 2024
  3. NeurIPS
    Gated Slot Attention for Efficient Linear-Time Sequence Modeling
    Yu Zhang*, Songlin Yang*, Ruijie Zhu, Yue Zhang, Leyang Cui, Yiqiao Wang, Bolun Wang, Freda Shi, Bailin Wang, Wei Bi, Peng Zhou, and Guohong Fu
    In , 2024
  4. ICML
    Gated Linear Attention Transformers with Hardware-Efficient Training
    Songlin Yang*, Bailin Wang*Yikang ShenRameswar Panda, and Yoon Kim
    In , 2024