Abstract
The purpose of this paper is to evaluate the possibility of using publicly available large language models for qualitative analysis of loosely structured textual data. An example of such data are the results of a survey containing “open-ended” questions. Using several publicly available large language models, the processing methodology of the results of the questionnaire, the subject of which was to determine the relationship between basic emotions and odours that evoke them, is demonstrated. The results of the data processing are in qualitative agreement with different large language models and the conclusions of the original methodology. Based on the results of this work, the possibilities and limitations of large language models as well as promising directions of their application are presented.