Work

Teaching Specific Scientific Knowledge into Large Language Models through Additional Training

Year: 2023

Type: preprint

Abstract: Through additional training, we explore embedding specialized scientific knowledge into the Llama 2 Large Language Model (LLM). Key findings reveal that effective knowledge integration requires reading texts from multiple perspectives, especially in instructional formats. We utilize text augmentation to tackle the scarcity of specialized texts, including style conversions and translations. Hyperparameter optimization proves crucial, with different size models (7b, 13b, and 70b) reasonably undergoing additional training. Validating our methods, we construct a dataset of 65,000 scientific papers. Although we have succeeded in partially embedding knowledge, the study highlights the complexities and limitations of incorporating specialized information into LLMs, suggesting areas for further improvement.

Source: arXiv (Cornell University)

DOI: 10.48550/arxiv.2312.03360

Authors: Hatakeyama-Sato, Kan, Igarashi, Yasuhiko, Katakami, Shun, Nabae, Yuta, Hayakawa, Teruaki

Institutions: Tokyo Institute of Technology, University of Tsukuba, The University of Tokyo

Language: English

Topic: Topic Modeling

Subfield: Artificial Intelligence

Field: Computer Science

Domain: Physical Sciences

SDG: Quality education

Open Access status: green

Funders: Ministry of Education, Culture, Sports, Science and Technology, Japan Society for the Promotion of Science