Selection of Large Language Model for development of Interactive Chat Bot for SaaS Solutions - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2024

Selection of Large Language Model for development of Interactive Chat Bot for SaaS Solutions

Résumé

Chat Bots play crucial role in modern world businesses. Development of a chatbot is a tedious and complex task that takes enormous amount of time. Therefore, to enable businesses to develop chatbots in short amount of time with small resources, we need to explore usage of LLMs for the chatbot development. In this article the researchers developed a prototype chatbot architecture that enables businesses to use LLMs interchangeably. Nonetheless, it notes the prototype's limitation in tracking conversation history, an area ripe for future enhancement. The current focus on pre-trained LLMs sets the stage for subsequent research into personalized, fine-tuned chatbot experiences for SaaS customers. There are dozens of LLMs existing already, therefore it is crucial to select the most capable LLM which will power the chatbot. It is crucial for the LLM to be cost-efficient, to be profitable for the business. This article evaluates three LLMs, endorsing ChatGPT for its superior speed, cost-effectiveness, and relevance, backed by OpenAI's pioneering status.
Fichier principal
Vignette du fichier
CoLinS_2024_cherednichenko.pdf (838.82 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-04545073 , version 1 (13-04-2024)

Licence

Paternité

Identifiants

  • HAL Id : hal-04545073 , version 1

Citer

Olga Cherednichenko, Dmytro Sytnikov, Nazarii Romankiv, Nataliia Sharonova, Polina Sytnikova. Selection of Large Language Model for development of Interactive Chat Bot for SaaS Solutions. 8th International Conference on Computational Linguistics and Intelligent Systems (CoLInS 2024), Apr 2024, Kharkiv, Ukraine. ⟨hal-04545073⟩
0 Consultations
0 Téléchargements

Partager

Gmail Facebook X LinkedIn More