Summary
In summary, GPT-4 outperforms the other two LLMs throughout the entire process, primarily due to the limitations of the models’ scale. Meanwhile, Qwen-72B performs considerably better than Qwen-14B and achieves comparable results, albeit slightly inferior, to those of GPT-4. We summarized our experiments in Table 2.
This is a preview of subscription content, log in via an institution to check access.
References
Pham C M, Hoyle A, Sun S, Iyyer M. Topicgpt: a prompt-based topic modeling framework. 2023, arXiv preprint arXiv: 2311.01449
Bai J, Bai S, Chu Y, et al. Qwen technical report. 2023, arXiv preprint arXiv: 2309.16609
Li Z, Zhang X, Zhang Y, Long D, Xie P, Zhang M. Towards general text embeddings with multi-stage contrastive learning. 2023, arXiv preprint arXiv: 2308.03281
McInnes L, Healy J, Saul N, Grossberger L. Umap: uniform manifold approximation and projection. The Journal of Open Source Software, 2018, 3(29): 861
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Competing interests The authors declare that they have no competing interests or financial conflicts to disclose.
Electronic supplementary material
Rights and permissions
About this article
Cite this article
Zeng, H., Sun, JM., Li, CS. et al. Foundation models for topic modeling: a case study. Front. Comput. Sci. 19, 192325 (2025). https://doi.org/10.1007/s11704-024-40069-7
Received:
Accepted:
Published:
DOI: https://doi.org/10.1007/s11704-024-40069-7