Large Language Models With Contrastive Decoding Algorithm for Hallucination Mitigation in Low‐Resource Languages

ABSTRACT Neural machine translation (NMT) has advanced with deep learning and large‐scale multilingual models, yet translating low‐resource languages often lacks sufficient training data and leads to hallucinations. This often results in translated content that diverges significantly from the source...

Descripció completa

Dades bibliogràfiques

Autors principals:	Zan Hongying, Arifa Javed, Muhammad Abdullah, Javed Rashid, Muhammad Faheem
Format:	Artigo
Idioma:	Inglês
Publicat:	Wiley 2025-08-01
Col·lecció:	CAAI Transactions on Intelligence Technology
Matèries:	ChatGLM contrastive decoding hallucination LLAMA LLM low resource NMT
Accés en línia:	https://doi.org/10.1049/cit2.70004
Etiquetes:	Afegir etiqueta Sense etiquetes, Sigues el primer a etiquetar aquest registre!