Krikri: Advancing Open Large Language Models for Greek

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Roussis, Dimitris, Voukoutis, Leon, Paraskevopoulos, Georgios, Sofianopoulos, Sokratis, Prokopidis, Prokopis, Papavasileiou, Vassilis, Katsamanis, Athanasios, Piperidis, Stelios, Katsouros, Vassilis
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866912404112146432
author Roussis, Dimitris
Voukoutis, Leon
Paraskevopoulos, Georgios
Sofianopoulos, Sokratis
Prokopidis, Prokopis
Papavasileiou, Vassilis
Katsamanis, Athanasios
Piperidis, Stelios
Katsouros, Vassilis
author_facet Roussis, Dimitris
Voukoutis, Leon
Paraskevopoulos, Georgios
Sofianopoulos, Sokratis
Prokopidis, Prokopis
Papavasileiou, Vassilis
Katsamanis, Athanasios
Piperidis, Stelios
Katsouros, Vassilis
contents We introduce Llama-Krikri-8B, a cutting-edge Large Language Model tailored for the Greek language, built on Meta's Llama 3.1-8B. Llama-Krikri-8B has been extensively trained on high-quality Greek data to ensure superior adaptation to linguistic nuances. With 8 billion parameters, it offers advanced capabilities while maintaining efficient computational performance. Llama-Krikri-8B supports both Modern Greek and English, and is also equipped to handle polytonic text and Ancient Greek. The chat version of Llama-Krikri-8B features a multi-stage post-training pipeline, utilizing both human and synthetic instruction and preference data, by applying techniques such as MAGPIE. In addition, for evaluation, we propose three novel public benchmarks for Greek. Our evaluation on existing as well as the proposed benchmarks shows notable improvements over comparable Greek and multilingual LLMs in both natural language understanding and generation as well as code generation.
format Preprint
id arxiv_https___arxiv_org_abs_2505_13772
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Krikri: Advancing Open Large Language Models for Greek
Roussis, Dimitris
Voukoutis, Leon
Paraskevopoulos, Georgios
Sofianopoulos, Sokratis
Prokopidis, Prokopis
Papavasileiou, Vassilis
Katsamanis, Athanasios
Piperidis, Stelios
Katsouros, Vassilis
Computation and Language
We introduce Llama-Krikri-8B, a cutting-edge Large Language Model tailored for the Greek language, built on Meta's Llama 3.1-8B. Llama-Krikri-8B has been extensively trained on high-quality Greek data to ensure superior adaptation to linguistic nuances. With 8 billion parameters, it offers advanced capabilities while maintaining efficient computational performance. Llama-Krikri-8B supports both Modern Greek and English, and is also equipped to handle polytonic text and Ancient Greek. The chat version of Llama-Krikri-8B features a multi-stage post-training pipeline, utilizing both human and synthetic instruction and preference data, by applying techniques such as MAGPIE. In addition, for evaluation, we propose three novel public benchmarks for Greek. Our evaluation on existing as well as the proposed benchmarks shows notable improvements over comparable Greek and multilingual LLMs in both natural language understanding and generation as well as code generation.
title Krikri: Advancing Open Large Language Models for Greek
topic Computation and Language
url https://arxiv.org/abs/2505.13772