Kakugo: Distillation of Low-Resource Languages into Small Language Models

Fuente: arXiv
Enregistré dans:
Détails bibliographiques
Auteurs principaux: Devine, Peter, Sanni, Mardhiyah, Adilazuarda, Farid, Loizaga, Julieta Gil, Haddow, Barry
Format: Preprint
Publié: 2026
Sujets:
Accès en ligne:
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866912835156574208
author Devine, Peter
Sanni, Mardhiyah
Adilazuarda, Farid
Loizaga, Julieta Gil
Haddow, Barry
author_facet Devine, Peter
Sanni, Mardhiyah
Adilazuarda, Farid
Loizaga, Julieta Gil
Haddow, Barry
contents We present Kakugo, a novel and cost-effective pipeline designed to train general-purpose Small Language Models (SLMs) for low-resource languages using only the language name as input. By using a large teacher model to generate synthetic prompts and translate instruction datasets, we produced training data and SLMs for 54 low-resource languages. Evaluations across a diverse set of general natural language processing tasks, including translation, classification, and question answering, demonstrate that our pipeline consistently improves performance over base models. With a total generation and training cost of under $50 per language, Kakugo offers an accessible method for communities to develop language-specific AI.
format Preprint
id arxiv_https___arxiv_org_abs_2601_14051
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Kakugo: Distillation of Low-Resource Languages into Small Language Models
Devine, Peter
Sanni, Mardhiyah
Adilazuarda, Farid
Loizaga, Julieta Gil
Haddow, Barry
Computation and Language
Artificial Intelligence
Machine Learning
We present Kakugo, a novel and cost-effective pipeline designed to train general-purpose Small Language Models (SLMs) for low-resource languages using only the language name as input. By using a large teacher model to generate synthetic prompts and translate instruction datasets, we produced training data and SLMs for 54 low-resource languages. Evaluations across a diverse set of general natural language processing tasks, including translation, classification, and question answering, demonstrate that our pipeline consistently improves performance over base models. With a total generation and training cost of under $50 per language, Kakugo offers an accessible method for communities to develop language-specific AI.
title Kakugo: Distillation of Low-Resource Languages into Small Language Models
topic Computation and Language
Artificial Intelligence
Machine Learning
url https://arxiv.org/abs/2601.14051