Conference paper · 2024 · Open access

Can LLMs Learn by Teaching for Better Reasoning? A Preliminary Study

Xuefei Ning, Zifu Wang, Shiyao Li, Shiyao Li, Peiran Yao, Tianyu Fu, Matthew B. Blaschko, Guohao Dai, Huazhong Yang, Yu Wang

neural information processing systems, 71188–71239 · Published 1 January 2024 · DOI 10.52202/079017-2275

Free PDFView at the publisherCite this paper

In brief

“Can LLMs Learn by Teaching for Better Reasoning? A Preliminary Study” is a 2024 conference paper by Xuefei Ning and 9 co-authors, published in neural information processing systems. OpenAlex counts 17 citations as of 15 November 2026, which puts it in the top 4% of works published in 2024 by citations. It is open access under a CC BY licence: the full text is free to read at arXiv (Cornell University).

Citations
17
Most cited in
2026 (8)
Field-weighted impact
2.60
References
30
Authors
10

Abstract

Teaching to improve student models (e.g., knowledge distillation) is an extensively studied methodology in LLMs. However, for humans, teaching improves not only students but also teachers, by fostering more rigorous and clear reasoning as well as knowledge building. We ask: Can LLMs also learn by teaching (LbT) for better reasoning? If the answer is yes, we can potentially unlock the possibility of continuously advancing the models without solely relying on human-produced data or stronger models. In this paper, we provide a preliminary exploration on this question. We show that LbT ideas can be incorporated into existing LLM training/prompting pipelines and bring improvements. Specifically, we design three methods, each mimicking one of the three levels of LbT: observing students' feedback, learning from the feedback, and learning iteratively, with the goals of improving answer accuracy without training or improving models' inherent capability with fine-tuning. We reveal some findings: (1) Teaching materials that make it easier for students to learn have clearer and more accurate logic when using in-context learning as the student's "learning" method; (2) Weak-to-strong generalization: LbT might help improve strong models by teaching weak models; (3) Diversity in students might help: teaching multiple students could be better than teaching one student or the teacher itself. We hope that our exploration can inspire future research on LbT and more broadly adopting the advanced techniques in education to improve LLMs. The code and website are at https://github.com/imagination-research/lbt and https://sites.google.com/view/llm-learning-by-teaching.

Cite this paper

Formatted from the catalogue record. Check names, capitals and page numbers against the paper before you submit.

APA 7

Ning, X., Wang, Z., Li, S., Li, S., Yao, P., Fu, T., Blaschko, M. B., Dai, G., Yang, H., & Wang, Y. (2024). Can LLMs Learn by Teaching for Better Reasoning? A Preliminary Study. neural information processing systems, 71188–71239. https://doi.org/10.52202/079017-2275

MLA 9

Ning, Xuefei, et al. “Can LLMs Learn by Teaching for Better Reasoning? A Preliminary Study.” neural information processing systems, 2024, pp. 71188–71239. https://doi.org/10.52202/079017-2275.

Chicago

Ning, Xuefei, Zifu Wang, Shiyao Li, Shiyao Li, Peiran Yao, Tianyu Fu, Matthew B. Blaschko, Guohao Dai, Huazhong Yang, and Yu Wang. 2024. “Can LLMs Learn by Teaching for Better Reasoning? A Preliminary Study.” neural information processing systems: 71188–71239. https://doi.org/10.52202/079017-2275.

Harvard

Ning, X. et al. (2024) ‘Can LLMs Learn by Teaching for Better Reasoning? A Preliminary Study’, neural information processing systems, pp. 71188–71239. doi:10.52202/079017-2275.

IEEE

X. Ning et al., “Can LLMs Learn by Teaching for Better Reasoning? A Preliminary Study,” neural information processing systems, pp. 71188–71239, 2024, doi: 10.52202/079017-2275.

BibTeX

@inproceedings{ning2024llms,
  title     = {{Can LLMs Learn by Teaching for Better Reasoning? A Preliminary Study}},
  author    = {Ning, Xuefei and Wang, Zifu and Li, Shiyao and Li, Shiyao and Yao, Peiran and Fu, Tianyu and Blaschko, Matthew B. and Dai, Guohao and Yang, Huazhong and Wang, Yu},
  booktitle = {neural information processing systems},
  year      = {2024},
  pages     = {71188--71239},
  doi       = {10.52202/079017-2275},
  url       = {https://doi.org/10.52202/079017-2275}
}

Download for a reference manager: BibTeX (.bib) · RIS for Zotero, Mendeley and EndNote (.ris)

Citations by year

YearCitationsShare of the best year
202444
202555
202688

Its field-weighted citation impact is 2.60: it has been cited 2.60 times as often as the average work of the same field, type and year (1.00 is the world average).

Topic and keywords

Physical Sciences › Computer Science › Artificial Intelligence › Topic Modeling

  • learning by teaching
  • large language models
  • knowledge distillation
  • student feedback
  • in-context learning
  • weak-to-strong generalization
  • student diversity
  • iterative learning

Work with this paper

Most-cited papers in Computer Science

All highly cited Computer Science papers

Questions about this paper

How do I cite this paper in APA style?

Ning, X., Wang, Z., Li, S., Li, S., Yao, P., Fu, T., Blaschko, M. B., Dai, G., Yang, H., & Wang, Y. (2024). Can LLMs Learn by Teaching for Better Reasoning? A Preliminary Study. neural information processing systems, 71188–71239. https://doi.org/10.52202/079017-2275

How many times has this paper been cited?

17 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 arXiv (Cornell University).

Who wrote this paper?

Xuefei Ning, Zifu Wang, Shiyao Li, Shiyao Li, Peiran Yao, Tianyu Fu, Matthew B. Blaschko, Guohao Dai, Huazhong Yang, Yu Wang.

What is the DOI of this paper?

10.52202/079017-2275. The permanent link is https://doi.org/10.52202/079017-2275.

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.