@inproceedings{yang-etal-2026-mid,
    title = "Mid-Think: Training-Free Intermediate-Budget Reasoning via Token-Level Triggers",
    author = "Yang, Van  and
      Wang, Shouren  and
      Ganguly, Debargha  and
      Li, Xinpeng  and
      Song, Chaoda  and
      Singh, Vikash  and
      Chaudhary, Vipin  and
      Han, Xiaotian",
    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.299/",
    pages = "6024--6038",
    ISBN = "979-8-89176-395-1",
    abstract = "Reasoning language models are controlled through explicit modes such as Think and No-think, yet we find that these behaviors are largely governed by a few token-level triggers rather than high-level instructions. Through attention analysis and controlled prompting experiments, we show that a leading ``Okay'' token induces reasoning behavior, while the newline pattern following `{\ensuremath{<}}/think{\ensuremath{>}}{`} suppresses it. Based on this observation, we propose Mid-Think, a simple training-free prompting format that combines these triggers to achieve intermediate-budget reasoning, consistently outperforming fixed-token and prompt-based baselines in terms of the accuracy{--}length trade-off. Furthermore, applying Mid-Think to RL training after SFT reduces training time by approximately 15{\%} while improving final performance of Qwen3-8B on AIME from 69.8{\%} to 72.4{\%} and on GPQA from 58.5{\%} to 61.1{\%}, demonstrating its effectiveness for both inference-time control and RL-based reasoning training."
}