@inproceedings{wang-etal-2026-reflect,
    title = "Reflect, Rewrite, Repeat: How Simple Arithmetic Enables Advanced Reasoning in Small Language Models",
    author = "Wang, Mengdie Flora  and
      Xie, Haochen  and
      Kim, Mun Young  and
      Chaudhury, Baishali  and
      Ashok, Meghana  and
      Gunturu, Suren  and
      Hong, Sungmin  and
      Woo, Jae Oh",
    editor = "Demberg, Vera  and
      Inui, Kentaro  and
      Marquez, Llu{\'i}s",
    booktitle = "Findings of the {A}ssociation for {C}omputational {L}inguistics: {EACL} 2026",
    month = mar,
    year = "2026",
    address = "Rabat, Morocco",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest-eacl/2026.findings-eacl.69/",
    pages = "1341--1363",
    ISBN = "979-8-89176-386-9",
    abstract = "Contemporary advancements in language model reasoning typically require computationally intensive reinforcement learning (RL) and massive datasets, creating barriers for resource-constrained teams. In this work, we demonstrate that high-quality, iterative training on minimal data can rival modern RL approaches. We introduce a resource-efficient framework that combines Direct Preference Optimization (DPO) and Supervised Fine-Tuning (SFT) with selective guidance from larger models, iteratively refining solutions through a ``reflect, rewrite, repeat'' cycle (R$^3$). Using Qwen 2.5 7B and Qwen 2.5 Math 7B as base models, our method shows meaningful performance improvements across arithmetic, symbolic and cognitive reasoning benchmarks{---}including GSM8K (83.1{\%} {\textrightarrow} 88.6{\%}), AIME{'}25@10 (20.0{\%} {\textrightarrow} 30.0{\%}) and LastLetterConcat (40.7{\%} {\textrightarrow} 53.3{\%}) problems. The model-agnostic nature of our R$^3$ framework is further demonstrated through substantial improvements when applied to Mistral and LLaMA-based models. Remarkably, these gains are achieved using mere 700 basic arithmetic training samples, in stark contrast to the hundreds of thousands of examples typically required by RL-based systems. Our results suggest that reasoning improvements need not strictly depend on large-scale data. By emphasizing strategically curated training grounded in foundational principles, we achieve competitive generalization with minimal resource overhead. Our R$^3$ pipeline also generates high-quality SFT data with high-fidelity reasoning traces as byproduct, further enabling scalable and annotation-free fine-tuning. Code is available.[{\ensuremath{<}}https://github.com/aws-samples/sample-for-reflect-rewrite-repeat{\ensuremath{>}}]"
}