The use of large language models in solving psychological reasoning problems

Authors

Chuganskaya A.

Annotation

The modern development of artificial intelligence has led to the active use of large language models (LLM) in various fields of scientific and professional knowledge. LLM is based on the principles of deep machine learning using a large amount of data and parameters aimed at solving problems in natural language. As part of the research on the use of large language models in scientific papers, we conducted a theoretical and analytical study based on materials from leading scientific journals on psychology in articles from 2023. Two trends of work were identified. In Russian works, the LLM model was used to develop expert categorization systems for socio‑ psychological phenomena. The use of Psychodiagnostic techniques for assessing the “abilities” of large language models and constructing reasoning is actively represented in foreign works. In LLM building systems, they often rely on psychological theories focused on building everyday knowledge, socially determined interactions, and systems of attributive processes. At the moment, there is a large expansion of work in the humanities using LLM, which makes it possible to develop new directions in the cognitive sciences, expand opportunities for implementation in robotics, but also carries many difficulties, primarily in assessing the quality of reasoning and the reliability of knowledge gained.

External links

Download proceedings (PDF) at eLibrary (in Russian, registration required): https://elibrary.ru/item.asp?id=89112922

Article's page at eLibrary: https://elibrary.ru/item.asp?id=91184086

Proceedings at the HSE website: https://publications.hse.ru/en/books/1212913875