Download file (3.33MB)
Download as zip (2.78MB)

コレクション

引用

説明:

Large language models (LLMs) are general-purpose tools with wide-ranging applications, including in materials science. In this work, we introduce aLLoyM, a fine-tuned LLM specifically trained on alloy compositions, temperatures, and their corresponding phase information. To develop aLLoyM, we curated question-and-answer (Q&A) pairs for binary and ternary phase diagrams using the open-source Computational Phase Diagram Database (CPDDB) and assessments based on CALPHAD (CALculation of PHAse Diagrams). We fine-tuned Mistral, an open-source pre-trained LLM, for two distinct Q&A formats: multiple-choice and short-answer. Benchmark evaluations demonstrate that fine-tuning substantially enhances performance on multiple-choice phase diagram questions. Moreover, the short-answer model of aLLoyM can generate novel phase diagrams from its components alone, suggesting that it may aid the discovery of new materials systems. To promote further research and adoption, we have publicly released the short-answer fine-tuned version of aLLoyM, along with the complete benchmarking Q&A dataset, on Hugging Face.

権利情報:

• Creative Commons BY-NC-ND Attribution-NonCommercial-NoDerivs 4.0 International
  

キーワード: LLM, phase diagram

刊行年月日: 2026-01-22

出版者: Springer Science and Business Media LLC

掲載誌:

• npj Computational Materials (ISSN: 20573960) vol. 12 97

研究助成金:

• Japan Society for the Promotion of Science 25KJ0870
• Japan Society for the Promotion of Science 25K01492
• MEXT Program: Data Creation and Utilization Type Material Research and Development Project
• Japan Science and Technology Agency,Japan JPMJCR21O2

原稿種別: 出版者版 (Version of record)

MDR DOI:

公開URL: https://doi.org/10.1038/s41524-026-01966-6

関連資料:

その他の識別子:

連絡先:

更新時刻: 2026-04-01 13:56:07 +0900

MDRでの公開時刻: 2026-04-01 16:26:13 +0900