Lightweight Prompt Adaptation via Self-Organizing Maps

• Samuel Henrique Santos Silva UPE
• Allan Miller Silva Lima UPE
• Denis Mayr Lima Martins USP
• Fernando Buarque de Lima Neto UPE

DOI: https://doi.org/10.5753/eniac.2026.30060

Resumo

Current methods for adapting and personalizing Large Language Models (LLMs) often force a choice between costly fine-tuning and unscalable manual prompt engineering. We propose Self-Organizing Prompt Maps (SOPM), which employs topological mapping to automate prompt adaptation. By mapping semantically related prompts onto a structured grid, SOPM enables contextually aware retrieval without additional LLM training. Our evaluation on two LLMs demonstrates that SOPM provides an interpretable way to navigate the prompt space while highlighting the importance of contrastive formulations for model-specific performance. SOPM provides a high-efficiency solution for production environments where low-overhead adaptation is paramount.

Referências

Arora, S., Narayan, A., Chen, M. F., Orr, L., Guha, N., Bhatia, K., Chami, I., and Re, C. (2023). Ask me anything: A simple strategy for prompting language models. In The Eleventh International Conference on Learning Representations.

Cheng, D., Huang, S., Bi, J., Zhan, Y., Liu, J., Wang, Y., Sun, H., Wei, F., Deng, W., and Zhang, Q. (2023). UPRISE: Universal prompt retrieval for improving zero-shot evaluation. In Bouamor, H., Pino, J., and Bali, K., editors, Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, pages 12318–12337, Singapore. Association for Computational Linguistics.

Cheng, W., Liu, J., Jiao, Z., Deng, J., and Ren, F. (2024). Ssp: A simple and safe prompt engineering for visual generation. In 2024 IEEE International Conference on Systems, Man, and Cybernetics (SMC), pages 4074–4079.

Chowdhury, F., Mihindukulasooriya, N., D’Souza, N. S., Samulowitz, H., Gupta, N., Hanusiak, T., and Kapitonow, M. (2026). Automatic prompt engineering with no task cues and no tuning.

Gao, Y., Xiong, Y., Gao, X., Jia, K., Pan, J., Bi, Y., Dai, Y., Sun, J., Wang, M., and Wang, H. (2024). Retrieval-augmented generation for large language models: A survey.

Gu, J. and Niu, H. (2024). Two-stage multi-modal prompt tuning for few-shot sentiment analysis. In 2024 IEEE International Conference on Systems, Man, and Cybernetics (SMC), pages 2856–2863.

Han, Z., Gao, C., Liu, J., Zhang, J., and Zhang, S. Q. (2024). Parameter-efficient fine-tuning for large models: A comprehensive survey.

Henderson, J., Merlo, P., Petroff, I., and Schneider, G. (2002). Using nlp to efficiently visualize text collections with soms. In Proceedings of the NLIS Workshop.

Hu, T. and Zhou, X.-H. (2024). Unveiling llm evaluation focused on metrics: Challenges and solutions.

Khacef, L., Miramond, B., Barrientos, D., and Upegui, A. (2019). Self-organizing neurons: toward brain-inspired unsupervised learning. In 2019 International Joint Conference on Neural Networks (IJCNN), pages 1–9.

Kohonen, T. (1990). The self-organizing map. Proceedings of the IEEE, 78(9):1464–1480.

Li, W., Wang, X., Li, W., and Jin, B. (2025). A survey of automatic prompt engineering: An optimization perspective.

Manduchi, L., Hüser, M., Vogt, J. E., Rätsch, G., and Horn, M. (2019). Deep probabilistic clustering with self-organizing maps. arXiv preprint arXiv:1910.01590.

Mao, W., Wu, J., Chen, W., Gao, C., Wang, X., and He, X. (2025). Reinforced prompt personalization for recommendation with large language models. ACM Trans. Inf. Syst., 43(3).

Papineni, K., Roukos, S., Ward, T., and Zhu, W.-J. (2002). Bleu: a method for automatic evaluation of machine translation. In Proceedings of the 40th Annual Meeting on Association for Computational Linguistics, ACL ’02, page 311–318, USA. Association for Computational Linguistics.

Tang, X., Zhai, H., Belwal, C., Thayanithi, V., Baumann, P., and Roy, Y. K. (2025). Dynamic context-aware prompt recommendation for domain-specific ai applications.

Xu, B., Wu, Y., Li, S., Song, H., Du, M., and Wang, H. (2024). A decomposition framework for class incremental text classification with dual prompt tuning. In 2024 International Joint Conference on Neural Networks (IJCNN), pages 1–8.

Zhang, X., Zhao, J., and LeCun, Y. (2015). Character-level convolutional networks for text classification. In Cortes, C., Lawrence, N., Lee, D., Sugiyama, M., and Garnett, R., editors, Advances in Neural Information Processing Systems, volume 28. Curran Associates, Inc.