https://stillkeeptry.github.io/Recent content on Kaitao SongHugoen-usSat, 30 May 2026 00:45:26 -0700https://stillkeeptry.github.io/about/Sat, 30 May 2026 00:45:26 -0700https://stillkeeptry.github.io/about/<p>Hi, I&rsquo;m <strong>Kaitao Song</strong>, currently serving as a Senior Research Scientist at Apple since May 2025. Previously, I was a Senior Researcher at Microsoft Research Asia since July 2021, focusing on AI Agents, Large Language Models, and related topics. Before that, I received my Bachelor and PhD degree at Nanjing University of Science and Technology in 2015 and 2021. My research interests include Foundation Models, Large Language Models, AI Agents, and NLP more broadly.</p>https://stillkeeptry.github.io/publications/Sat, 30 May 2026 00:45:26 -0700https://stillkeeptry.github.io/publications/<p style="color: var(--secondary); font-size: 0.92rem; margin-top: -0.5rem;"> * = equal contribution &nbsp;&middot;&nbsp; # = corresponding author. See also my <a href="https://scholar.google.com/citations?user=LLk9dR8AAAAJ&hl=en">Google Scholar</a>. </p> <h2 id="foundation-model">Foundation Model</h2> <ul> <li> <p><strong>Chain-of-Model Learning for Language Model</strong>.<br> <strong>Kaitao Song</strong>*^,#, Xiaohua Wang*, Xu Tan*, Huiqiang Jiang, Chengruidong Zhang, Yongliang Shen, Cen Lu, Zihao Li, Zifan Song, Caihua Shan, Yansen Wang, Kan Ren, Xiaoqing Zheng, Tao Qin, Yuqing Yang, Dongsheng Li, Lili Qiu.<br> <em>NeurIPS 2025.</em></p> </li> <li> <p><strong>Let LLMs Break Free from Overthinking via Self-Braking Tuning</strong>.<br> Haoran Zhao, Yuchen Yan, Yongliang Shen, Haolei Xu, Wenqi Zhang, <strong>Kaitao Song</strong>, Jian Shao, Weiming Lu, Jun Xiao, Yueting Zhuang.<br> <em>NeurIPS 2025.</em></p>