@inproceedings{yang-etal-2026-chain,
    title = "How Chain-of-Thought Works? Tracing Information Flow from Decoding, Projection, and Activation",
    author = "Yang, Hao  and
      Zhao, Qinghua  and
      Li, Lei  and
      Meng, Lingyi  and
      Yu, Mengda",
    editor = "Liakata, Maria  and
      Moreira, Viviane P.  and
      Zhang, Jiajun  and
      Jurgens, David",
    booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {ACL} 2026",
    month = jul,
    year = "2026",
    address = "San Diego, California, United States",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest-acl/2026.findings-acl.255/",
    pages = "5166--5199",
    ISBN = "979-8-89176-395-1",
    abstract = "Chain-of-Thought (CoT) prompting significantly enhances model reasoning, yet its internal mechanisms remain poorly understood. We analyze CoT{'}s operational principles by reversely tracing information flow across decoding, projection, and activation phases. Our quantitative analysis suggests that CoT may serve as a decoding space pruner, leveraging answer templates to guide output generation, with higher template adherence strongly correlating with improved performance. Furthermore, we surprisingly find that CoT modulates neuron engagement in a task-dependent manner: reducing neuron activation in open-domain tasks, yet increasing it in closed-domain scenarios. These findings offer a novel mechanistic interpretability framework and critical insights for the NLP community, enabling targeted CoT interventions to design more efficient and robust prompts. We released our code and data at https://github.com/How-Young-X/cot"
}