Gusching, M. Jonathan (2023) Gros modèles de langage et connaissances : état de l’art, enjeux, applications PFE - Project Graduation, ENSTA.

[img]
Preview
PDF
766Kb

Abstract

Since the emergence of ChatGPT in 2022, large language models have experienced a surge in popularity and increased usage in various fields. The ability to harness vast amounts of textual information is rooted in a broader enthusiasm for generative models. The goal of this internship and, therefore, this report is, first and foremost, to provide an overview of the state of the art in terms of large language models, the challenges they pose, their potential, and trends. Additionally, we will outline the possibilities for application in journalism, given that few publications have been written on this topic and this field of application is relatively under-documented. Finally, we will conclude with the implementation of a web interface for experimenting with large language models for journalistic purposes and a discussion of the advantages and disadvantages of such applications. Lastly, we will introduce the publication resulting from the internship as part of the evaluation of language model knowledge.

Item Type:Thesis (PFE - Project Graduation)
Uncontrolled Keywords:Large language models have been suffering from the hallucination phenomenon, that is, writing false information, but in such a confident style and grammar that it sounds believably true. This roots in different causes, one being the lack of up-to-date inf
Subjects:Information and Communication Sciences and Technologies
ID Code:9823
Deposited By:Jonathan Gusching
Deposited On:06 oct. 2023 17:12
Dernière modification:06 oct. 2023 17:12

Repository Staff Only: item control page