Download pdf
GRPO and Reflection Reward for Mathematical Reasoning in Large Language Models
- 1 School of Computer Science and Engineering, Sun Yat-sen University(SYSU), Guangzhou, Guangdong Province, 510006, China
* Author to whom correspondence should be addressed.
Abstract
The enhancement of reasoning capabilities in large language models (LLMs) has garnered significant attention, with supervised fine-tuning (SFT) and reinforcement learning emerging as dominant paradigms. While recent studies recognize the importance of reflection in reasoning processes, existing methodologies seldom address proactive reflection encouragement during training. This study focuses on mathematical reasoning by proposing a four-stage framework integrating Group Relative Policy Optimization (GRPO) with the reflection reward mechanisms to strengthen LLMs' self-reflective capabilities. Besides, this approach incorporates established accuracy and format reward. Experimental results demonstrate GRPO's state-of-the-art performance through reflection-encouraged training, with ablation studies confirming the reflection reward's pivotal role. Comparative evaluations demonstrate full-parameter SFT's superiority over low-rank adaptation (LoRA) despite heightened computational demands. Building on these cumulative findings, this research substantiates GRPO's methodological significance in post-training optimization and envisions its potential to serve as a pivotal enabler for future LLM-based intelligent agents through the synergistic integration of cognitive rewards with dynamic environmental interactions.
Keywords
large language models (LLMs), Group Relative Policy Optimization (GRPO), reflection, reasoning
[1]. Kumar, K., Ashraf, T., Thawakar, O., Anwer, R. M., Cholakkal, H., Shah, M., ... & Khan, F. S. (2025). Llm post-training: A deep dive into reasoning large language models. arXiv preprint arXiv:2502.21321.
[2]. Chu, T., Zhai, Y., Yang, J., Tong, S., Xie, S., Schuurmans, D., ... & Ma, Y. (2025). Sft memorizes, rl generalizes: A comparative study of foundation model post-training. arXiv preprint arXiv:2501.17161.
[3]. Shao, Z., Wang, P., Zhu, Q., Xu, R., Song, J., Bi, X., ... & Guo, D. (2024). Deepseekmath: Pushing the limits of mathematical reasoning in open language models. arXiv preprint arXiv:2402.03300.
[4]. Guo, D., Yang, D., Zhang, H., Song, J., Zhang, R., Xu, R., ... & He, Y. (2025). Deepseek-r1: Incentivizing reasoning capability in llms via reinforcement learning. arXiv preprint arXiv:2501.12948.
[5]. Hu, E. J., Shen, Y., Wallis, P., Allen-Zhu, Z., Li, Y., Wang, S., ... & Chen, W. (2022). Lora: Low-rank adaptation of large language models. ICLR, 1(2), 3.
[6]. Open R1: A fully open reproduction of DeepSeek-R1. Retrieved 4/9/2025, https://github.com/huggingface/open-r1
[7]. Red-Scarff. (2025). GRPO_reflection: Thinking_LLM. GitHub. Retrieved 4/9/2025, https://github.com/Red-Scarff/GRPO_reflection.git
[8]. Open R1. (n.d.). OpenR1-Math-220k. Hugging Face. Retrieved 4/9/2025, https://huggingface.co/datasets/open-r1/OpenR1-Math-220k
[9]. Digital Learning GmbH. (n.d.). MATH-lighteval. Hugging Face. Retrieved 4/9/2025, https://huggingface.co/datasets/DigitalLearningGmbH/MATH-lighteval
[10]. Jia, M. (2024). AIME_2024. Hugging Face. Retrieved 4/9/2025, https://huggingface.co/datasets/Maxwell-Jia/AIME_2024
[11]. Hendrycks, D., Burns, C., Kadavath, S., Arora, A., Basart, S., Tang, E., ... & Steinhardt, J. (2021). Measuring mathematical problem solving with the math dataset. arXiv preprint arXiv:2103.03874.
[12]. Rein, D., Hou, B. L., Stickland, A. C., Petty, J., Pang, R. Y., Dirani, J., ... & Bowman, S. R. (2024). Gpqa: A graduate-level google-proof q&a benchmark. In First Conference on Language Modeling.
[13]. Yang, A., Zhang, B., Hui, B., Gao, B., Yu, B., Li, C., ... & Zhang, Z. (2024). Qwen2. 5-math technical report: Toward mathematical expert model via self-improvement. arXiv preprint arXiv:2409.12122.
[14]. Raffel, C., Shazeer, N., Roberts, A., Lee, K., Narang, S., Matena, M., ... & Liu, P. J. (2020). Exploring the limits of transfer learning with a unified text-to-text transformer. Journal of machine learning research, 21(140), 1-67.
[15]. Dettmers, T., Lewis, M., Belkada, Y., & Zettlemoyer, L. (2022). Gpt3. int8 (): 8-bit matrix multiplication for transformers at scale. Advances in neural information processing systems, 35, 30318-30332.
[16]. Zeng, W., Huang, Y., Liu, W., He, K., Liu, Q., Ma, Z., & He, J. 7b model and 8k examples: Emerging reasoning with reinforcement learning is both effective and efficient.
[17]. Asai, A., Wu, Z., Wang, Y., Sil, A., & Hajishirzi, H. (2023, October). Self-rag: Learning to retrieve, generate, and critique through self-reflection. In The Twelfth International Conference on Learning Representations.
[18]. Zeng, W., Huang, Y., Liu, W., He, K., Liu, Q., Ma, Z., & He, J. 7b model and 8k examples: Emerging reasoning with reinforcement learning is both effective and efficient.
[19]. Gandhi, K., Chakravarthy, A., Singh, A., Lile, N., & Goodman, N. D. (2025). Cognitive Behaviors that Enable Self-Improving Reasoners, or, Four Habits of Highly Effective STaRs. arXiv preprint arXiv:2503.01307.
Cite this article
Wang,Z. (2025). GRPO and Reflection Reward for Mathematical Reasoning in Large Language Models. Applied and Computational Engineering,154,161-166.
Data availability
The datasets used and/or analyzed during the current study will be available from the authors upon reasonable request.
Disclaimer/Publisher's Note
The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of EWA Publishing and/or the editor(s). EWA Publishing and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.
About volume
Volume title: Proceedings of CONF-SEML 2025 Symposium: Machine Learning Theory and Applications
© 2024 by the author(s). Licensee EWA Publishing, Oxford, UK. This article is an open access article distributed under the terms and
conditions of the Creative Commons Attribution (CC BY) license. Authors who
publish this series agree to the following terms:
1. Authors retain copyright and grant the series right of first publication with the work simultaneously licensed under a Creative Commons
Attribution License that allows others to share the work with an acknowledgment of the work's authorship and initial publication in this
series.
2. Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the series's published
version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgment of its initial
publication in this series.
3. Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and
during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work (See
Open access policy for details).