Conference paper · 2026 · Open access
Post-Training for Reasoning LLMs with Reinforcement Learning: A Stability–Efficiency Perspective
International Conference on Automation and Computing (ICAC), 1–6 · Published 1 August 2026 · DOI 10.1109/icac70325.2026.11712300
In brief
“Post-Training for Reasoning LLMs with Reinforcement Learning: A Stability–Efficiency Perspective” is a 2026 conference paper by Liuhaichen Yang and 7 co-authors, published in International Conference on Automation and Computing (ICAC). OpenAlex counts 0 citations as of 15 November 2026. It is open access under a CC BY licence: the full text is free to read at International Conference on Automation and Computing (ICAC).
- Citations
- 0
- References
- 8
- Authors
- 8
Abstract
Reinforcement Learning (RL)-based post-training has become a key mechanism for improving reasoning Large Language Models (LLMs), especially when automated verifiers enable Reinforcement Learning with Verifiable Reward (RLVR). Recent methods, however, are increasingly modular: many apparent algorithmic advances recombine a small set of recurring components, including group-relative baselines, trust-region control, dense or process-level feedback, replay, and throughput-oriented data routing. This survey organises this literature through a unified design map with three axes: learning-signal source and density, optimisation regime and trust-region control, and structural granularity and agentic scope. We further use a compact method-by-motif view to expose common stability–efficiency trade-offs across representative methods. Rather than treating stability as a formal convergence guarantee, we focus on auditable failure signatures, including reward degeneracy, ratio-tail growth, entropy collapse, length bias, and off-policy mismatch. We conclude with practical evaluation guidance based on matched compute, controlled length, and diagnostic reporting, and identify open problems in dense-signal causality, critic-free off-policy RL, and long-horizon credit assignment.
Cite this paper
Formatted from the catalogue record. Check names, capitals and page numbers against the paper before you submit.
APA 7
Yang, L., Zhu, H., Zhong, Z., Bai, N., Tan, X., Yin, Q., Yan, Z., & Tang, Z. (2026). Post-Training for Reasoning LLMs with Reinforcement Learning: A Stability–Efficiency Perspective. International Conference on Automation and Computing (ICAC), 1–6. https://doi.org/10.1109/icac70325.2026.11712300
MLA 9
Yang, Liuhaichen, et al. “Post-Training for Reasoning LLMs with Reinforcement Learning: A Stability–Efficiency Perspective.” International Conference on Automation and Computing (ICAC), 2026, pp. 1–6. https://doi.org/10.1109/icac70325.2026.11712300.
Chicago
Yang, Liuhaichen, Hanshang Zhu, Zhengyang Zhong, Ningwei Bai, Xinyu Tan, Qichen Yin, Zonghao Yan, and Zezhi Tang. 2026. “Post-Training for Reasoning LLMs with Reinforcement Learning: A Stability–Efficiency Perspective.” International Conference on Automation and Computing (ICAC): 1–6. https://doi.org/10.1109/icac70325.2026.11712300.
Harvard
Yang, L. et al. (2026) ‘Post-Training for Reasoning LLMs with Reinforcement Learning: A Stability–Efficiency Perspective’, International Conference on Automation and Computing (ICAC), pp. 1–6. doi:10.1109/icac70325.2026.11712300.
IEEE
L. Yang et al., “Post-Training for Reasoning LLMs with Reinforcement Learning: A Stability–Efficiency Perspective,” International Conference on Automation and Computing (ICAC), pp. 1–6, 2026, doi: 10.1109/icac70325.2026.11712300.
BibTeX
@inproceedings{yang2026post,
title = {{Post-Training for Reasoning LLMs with Reinforcement Learning: A Stability–Efficiency Perspective}},
author = {Yang, Liuhaichen and Zhu, Hanshang and Zhong, Zhengyang and Bai, Ningwei and Tan, Xinyu and Yin, Qichen and Yan, Zonghao and Tang, Zezhi},
booktitle = {International Conference on Automation and Computing (ICAC)},
year = {2026},
pages = {1--6},
doi = {10.1109/icac70325.2026.11712300},
url = {https://doi.org/10.1109/icac70325.2026.11712300}
}
Download for a reference manager: BibTeX (.bib) · RIS for Zotero, Mendeley and EndNote (.ris)
Topic and keywords
Physical Sciences › Computer Science › Artificial Intelligence › Reinforcement Learning in Robotics
- reasoning LLMs
- reinforcement learning with verifiable rewards
- entropy collapse
- replay
- data routing
Work with this paper
Most-cited papers in Computer Science
- ImageNet classification with deep convolutional neural networksAlex Krizhevsky, Ilya Sutskever, Geoffrey E. Hinton · Communications of the ACM · 2017 · 114,726 citations
- Long Short-Term MemorySepp Hochreiter, Jürgen Schmidhuber · Neural Computation · 1997 · 109,097 citations
- A Mathematical Theory of CommunicationClaude E. Shannon · Bell System Technical Journal · 1948 · 91,736 citations
- Deep learningYann LeCun, Yoshua Bengio, Geoffrey E. Hinton · Nature · 2015 · 90,226 citations
- Faster R-CNN: Towards Real-Time Object Detection with Region Proposal NetworksShaoqing Ren, Kaiming He, Ross Girshick and 1 more · IEEE Transactions on Pattern Analysis and Machine Intelligence · 2016 · 74,143 citations
Questions about this paper
How do I cite this paper in APA style?
Yang, L., Zhu, H., Zhong, Z., Bai, N., Tan, X., Yin, Q., Yan, Z., & Tang, Z. (2026). Post-Training for Reasoning LLMs with Reinforcement Learning: A Stability–Efficiency Perspective. International Conference on Automation and Computing (ICAC), 1–6. https://doi.org/10.1109/icac70325.2026.11712300
How many times has this paper been cited?
0 times, according to OpenAlex on 15 November 2026. Google Scholar and Scopus count different sets of sources, so their numbers differ.
Is this paper free to read?
Yes. It is open access: read the free full text at International Conference on Automation and Computing (ICAC).
Who wrote this paper?
Liuhaichen Yang, Hanshang Zhu, Zhengyang Zhong, Ningwei Bai, Xinyu Tan, Qichen Yin, Zonghao Yan, Zezhi Tang.
What is the DOI of this paper?
10.1109/icac70325.2026.11712300. The permanent link is https://doi.org/10.1109/icac70325.2026.11712300.
Bibliographic data from OpenAlex (CC0), last checked 15 November 2026. ResearchGuru does not host this paper; the links go to the publisher or the repository that does.