70B-parameter large language models in Japanese medical question-answering

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Sukeda, Issey, Kishikawa, Risa, Kodera, Satoshi
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916296344469504
author Sukeda, Issey
Kishikawa, Risa
Kodera, Satoshi
author_facet Sukeda, Issey
Kishikawa, Risa
Kodera, Satoshi
contents Since the rise of large language models (LLMs), the domain adaptation has been one of the hot topics in various domains. Many medical LLMs trained with English medical dataset have made public recently. However, Japanese LLMs in medical domain still lack its research. Here we utilize multiple 70B-parameter LLMs for the first time and show that instruction tuning using Japanese medical question-answering dataset significantly improves the ability of Japanese LLMs to solve Japanese medical license exams, surpassing 50\% in accuracy. In particular, the Japanese-centric models exhibit a more significant leap in improvement through instruction tuning compared to their English-centric counterparts. This underscores the importance of continual pretraining and the adjustment of the tokenizer in our local language. We also examine two slightly different prompt formats, resulting in non-negligible performance improvement.
format Preprint
id arxiv_https___arxiv_org_abs_2406_14882
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle 70B-parameter large language models in Japanese medical question-answering
Sukeda, Issey
Kishikawa, Risa
Kodera, Satoshi
Computation and Language
Since the rise of large language models (LLMs), the domain adaptation has been one of the hot topics in various domains. Many medical LLMs trained with English medical dataset have made public recently. However, Japanese LLMs in medical domain still lack its research. Here we utilize multiple 70B-parameter LLMs for the first time and show that instruction tuning using Japanese medical question-answering dataset significantly improves the ability of Japanese LLMs to solve Japanese medical license exams, surpassing 50\% in accuracy. In particular, the Japanese-centric models exhibit a more significant leap in improvement through instruction tuning compared to their English-centric counterparts. This underscores the importance of continual pretraining and the adjustment of the tokenizer in our local language. We also examine two slightly different prompt formats, resulting in non-negligible performance improvement.
title 70B-parameter large language models in Japanese medical question-answering
topic Computation and Language
url https://arxiv.org/abs/2406.14882