22 janvier 2024
http://creativecommons.org/licenses/by/ , info:eu-repo/semantics/OpenAccess
Jean Barré, « Large Language models in quest for adventure », HAL SHS (Sciences de l’Homme et de la Société), ID : 10670/1.65c5fa...
This study is a methodological effort on the use of generative LLMs in a literary context. We attempted to identify what constitutes stereotypical passages in adventure novels. To achieve this, we leveraged LLMs as a tool for large-scale annotation. Human evaluation of the performance shows a convergence of over nine times out of ten with the automatic annotations. We utilized these annotations to train a model capable of recognizing adventure passages with 86% accuracy. Upon generalization to a large corpus of literary texts, we discuss the results and their potential implications.