Xiangchen Song (宋祥辰)

xiangchen-head-2024.jpeg

I am a PhD student in Machine Learning Department at Carnegie Mellon University, fortunately advised by Prof. Kun Zhang (CMU-CLeaR Group). Before that, I was an undergraduate student in Data Mining Group in Computer Science at University of Illinois at Urbana-Champaign, where I was advised by Prof. Jiawei Han.

research

My research lies at the intersection of Machine Learning and Causality, with a focus on interpretability through provable causal representations.

I develop causal representation learning methods for sequential data—including
time series (NeurIPS’23, NeurIPS’24, arXiv’24, ICLR’25), video (ICML’24, NeurIPS’24), and text (NeurIPS’25). More recently, I am particularly interested in extending this causal lens to study large language models through mechanistic interpretability, aiming to both understand (NeurIPS’25, MechInterp@NeurIPS’25) and control (ICML’25) their latent reasoning processes toward more reliable and principled way.

contact

The easiest way to reach me is email. My address is xiangchs [at] cs [dot] cmu [dot] edu.

news

Sep 23, 2025 Two papers about efficient LLM reasoning have been accepted to NeurIPS 2025 Workshop on Efficient Reasoning (ER@NeurIPS’2025)!!
Sep 22, 2025 Two papers on mechanistic interpretability have been accepted to Mechanistic Interpretability Workshop at NeurIPS 2025 (MechInterp@NeurIPS’2025)!!
Sep 18, 2025 One paper “LLM Interpretability with Identifiable Temporal-Instantaneous Representation” has been accepted to The Thirty-ninth Conference on Neural Information Processing Systems (NeurIPS’2025)!!
May 01, 2025 One paper “Reflection-Window Decoding: Text Generation with Selective Refinement” been accepted to The Forty-Second International Conference on Machine Learning (ICML’2025)!!
Jan 22, 2025 One paper “On the Identification of Temporal Causal Representation with Instantaneous Dependence” has been accepted to The Thirteenth International Conference on Learning Representations (ICLR’2025) with oral presentation!!

selected publications

  1. LLM Interpretability with Identifiable Temporal-Instantaneous Representation
    Xiangchen Song*, Jiaqi Sun*, Zijian Li, Yujia Zheng, and Kun Zhang
    In The Thirty-ninth Annual Conference on Neural Information Processing Systems, Dec 2025
  2. Position: Mechanistic Interpretability Should Prioritize Feature Consistency in SAEs
    In Mechanistic Interpretability Workshop at NeurIPS (Spotlight), Dec 2025
  3. Causal Temporal Representation Learning with Nonstationary Sparse Transition
    In The Thirty-eighth Annual Conference on Neural Information Processing Systems, Dec 2024
  4. Temporally Disentangled Representation Learning under Unknown Nonstationarity
    In Thirty-seventh Conference on Neural Information Processing Systems, Dec 2023