Release:
2025. Vol. 11. № 1 (41)About the authors:
Svetlana V. Ostapenko, Applicant at the Department of Media Communications, Advertising Technologies and Public Relations, Altai State University, Barnaul, RussiaAbstract:
The increasing importance of formal text parameterization in internet communication has led to the popularization of large language models (LLMs) capable of generating natural language texts based on user-defined parameters through a prompt interface. Prompt engineering as a technique for improving user-program interaction by modeling relevant queries has emerged as a valuable tool for enhancing text generation mechanisms. This paper explores the phenomenon of linguistic optimization and its impact on the quality of the text. The research aims to analyze the functioning of text generation queries, investigate the possibilities of formalization and parameterization of the prompt, and experimentally study the impact of prompt detalization on the quality of the generated text. The results show that excessive prompt parameterization generally deteriorates the lexical diversity of the text, indicating model overfitting. Optimal prompt parameters for generating high-quality text vary within two to three parameters (construct type, indication of text genre and style parameters), suggesting the need for prompt formalization to achieve better text generation results. Applying prompt formalization for specifying text generation tasks links prompt engineering with cognitive engineering of knowledge. The prompt that predicts text content serves as a cognitive equivalent of the speech event model, influenced by metatextual data about the communication task available to the user. The prospects of prompt engineering research are not limited to improving artificial text creation processes, but also involve the development of structured approaches to the ontologizing of knowledge paradigms represented in texts of a particular discourse.Keywords:
References:
Wierzbicka, A. (1985). Speech acts. In N. D. Aruutynova & E. V. Paducheva (Eds.), New in Foreign Linguistics (Vol. 16. Linguistic Pragmatics, pp. 251–275). Progress. [In Russian]
Zakharova, E. Yu., Savina, O. Yu. (2020). Lexical diversity of text and its measurement methods. Tyumen State University Herald. Humanities Research. Humanitates, 6(1), 20–34. https://doi.org/10.21684/2411-197X-2020-6-1-20-34 [In Russian]
Nikolaev, F. A. (2018). Structural and semantic, and communicative-pragmatic characteristics of English-language advertising SEO texts [Cand. Sc. (Philology) dissertation, SPbGU]. [In Russian]
Poloyan, A. V., & Pridod’ko, K. D. (2023). Composition of a prompt in the midjourney neural network for generating illustrations for the works of A. P. Chekhov. In M. A. Lapteva (Ed.), Informatsionnye tekhnologii v gumanitarnykh issledovaniiah: materialy Mezhdunarodnoy nauchno-prakticheskoy konferentsii. Krasnoyarsk, 25–28 sentyabrya 2023 g. [Information Technologies in Humanities: Proceedings of the International Research Conference. Krasnoyarsk, Sep. 25–28, 2023] (pp. 194–199). Sibirskiy federalniy universitet. [In Russian]
Sokolova, M. E. (2024). ChatGPT and prompt engineering: prospects for integrating generative neural networks into science. Naukovedcheskie issledovaniia, (1), 92–109. https://doi.org/10.31249/scis/2024.01.07 [In Russian]
Soldatkina, Ya. V., & Chernavskii, A. S. (2023). Generative language models as a crucial phenomenon of media culture at the beginning of the XXI century. Nauka i shkola, 4, 44–56. http://doi.org/10.31862/1819-463X-2023-4-44-56 [In Russian]
Halina, N. V. (2014). Linguistic engineering and linguistic of engineering: adaptation of linguistic paradigm for circumstance of engineering epoch. Koncept, 20, 3811–3815. https://e-koncept.ru/2014/55027.htm [In Russian]
Cherkasova, M. N., & Taktarova, A. V. (2024). Artificially generated academic text (а linguopragmatic aspect). Philology. Theory & Practice, 17(7), 2551–2557. https://doi.org/10.30853/phil20240363 [In Russian]
Shelestiuk, E. V., & Shchetinkina, E. A. (2023). Stochasticity and entropy in linguistics. Vestnik Chelyabinskogo gosudarstvennogo universiteta, (2), 150–165. [In Russian]
Johansson, V. (2008). Lexical diversity and lexical density in speech and writing: a developmental perspective. Working Papers, 53, 61–79. Retrieved Jan. 5, 2025, from https://journals.lub.lu.se/LWPL/article/view/2273
Koizumi, R. (2012). Relationships between text length and lexical diversity measures: Can we use short texts of less than 100 tokens? Vocabulary Learning and Instruction, 1(1), 60–69. https://doi.org/10.7820/vli.v01.1.koizumi
McCarthy, P. M., & Jarvis, S. (2010). MTLD, vocD-D, and HD-D: a validation study of sophisticated approaches to lexical diversity assessment. Behavior Research Methods, 42(2), 381–392. https://doi.org/10.3758/brm.42.2.381
Ure, J. (1971). Lexical density and register differentiation. In G. Perren and J. L. M. Trim (Eds), Applications of Linguistics: Selected Papers of the Second International Congress of Applied Linguistics, Cambridge 1969 (pp. 443–452). Cambridge University Press.
White, J., Fu, Q., Hays, S., Sandborn, M., Olea, C., Gilbert, H., Elnashar, A., Spencer-Smith, J., & Schmidt, D. C. (2023). A prompt pattern catalog to enhance prompt engineering with ChatGPT. arXiv. https://doi.org/10.48550/arXiv.2302.11382
