Too Late to Train, Too Early To Use? A Study on Necessity and Viability of Low-Resource Bengali LLMs

Each new generation of English-oriented Large Language Models (LLMs) exhibits enhanced cross-lingual transfer capabilities and significantly outperforms older LLMs on low-resource languages. This prompts the question: Is there a need for LLMs dedicated to a particular low-resource language? We aim t...

Ausführliche Beschreibung

Bibliographische Detailangaben

Veröffentlicht in:	arXiv.org 2024-12
Hauptverfasser:	Mahfuz, Tamzeed, Dey, Satak Kumar, Naswan, Ruwad, Hasnaen Adil, Khondker, Salman Sayeed, Shahgir, Haz Sameen
Format:	Artikel
Sprache:	eng
Schlagworte:	Datasets Effectiveness Encoders-Decoders Large language models Machine translation Natural language processing Performance degradation Questions
Online-Zugang:	Volltext
Tags:	Tag hinzufügen Keine Tags, Fügen Sie den ersten Tag hinzu!