@inproceedings{ge-etal-2026-ada,
    title = "{A}da-{RS}: Adaptive Rejection Sampling for Selective Thinking",
    author = "Ge, Yirou  and
      Li, Yixi  and
      Chiu, Alec M.  and
      Shekhar, Shivani  and
      Pan, Zijie  and
      Thangali, Avinash  and
      Chuang, Yun-Shiuan  and
      Kulkarni, Chaitanya  and
      Kona, Uma  and
      Pang, Linsey  and
      Mehrotra, Prakhar",
    editor = "Li, Yunyao  and
      Rehm, Georg  and
      Tu, Mei",
    booktitle = "Proceedings of the 64th Annual Meeting of the {A}ssociation for {C}omputational {L}inguistics ({ACL} 2026)",
    month = jul,
    year = "2026",
    address = "San Diego, California, USA",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest-acl/2026.acl-industry.88/",
    pages = "1265--1276",
    ISBN = "979-8-89176-394-4",
    abstract = "Large language models (LLMs) are increasingly being deployed in cost- and latency-sensitive settings. While chain-of-thought improves reasoning, it can waste tokens on simple requests. We study selective thinking for tool-using LLMs and introduce Adaptive Rejection Sampling (Ada-RS), an algorithm-agnostic sample filtering framework for learning selective and efficient reasoning. For each given context, Ada-RS scores multiple sampled completions with an adaptive length-penalized reward then applies stochastic rejection sampling to retain only high-reward candidates (or preference pairs) for downstream optimization. We demonstrate how Ada-RS plugs into both preference pair (e.g. DPO) or grouped policy optimization strategies (e.g. DAPO). Using Qwen3-8B with LoRA on a synthetic tool call-oriented e-commerce benchmark, Ada-RS improves the accuracy-efficiency frontier over standard algorithms by reducing average output tokens by up to {\ensuremath{\sim}}80{\%} and reducing thinking rate by up to {\ensuremath{\sim}}95{\%} while maintaining or improving tool call accuracy. We further demonstrate that these gains generalize across model scales (Qwen3-1.7B, 8B, 14B) and domains ({\ensuremath{\tau}} 2-Bench airline and telecom). These results highlight that training signal selection is a powerful lever for efficient reasoning in latency-sensitive deployments."
}