XiHeFusion: Harnessing Large Language Models for Science Communication in Nuclear Fusion

2025-02-08

AI Summary

To lower the barrier to understanding nuclear fusion science and promote public engagement and scientific literacy, this work introduces XiHeFusion—the first large language model (LLM) specifically designed for nuclear fusion outreach. Built upon Qwen2.5-14B, it undergoes supervised fine-tuning on a diverse, multi-source knowledge corpus comprising arXiv preprints, electronic textbooks, theses, and web-scraped fusion-related materials, augmented with chain-of-thought (CoT) prompting to enhance physical reasoning and interpretability. We propose a novel, comprehensive evaluation benchmark—FusionSciBench—covering 180+ questions spanning conceptual understanding, quantitative reasoning, and real-world applications in fusion science communication. Experimental results demonstrate that XiHeFusion significantly outperforms general-purpose LLMs on fusion-specific科普 QA tasks. The model and benchmark are fully open-sourced to foster community-driven development, validation, and deployment in science education and outreach.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: (Large) Language Models
• Humans and AI: Explainable AI (XAI) for Human Understanding

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

Nuclear fusion is one of the most promising ways for humans to obtain infinite energy. Currently, with the rapid development of artificial intelligence, the mission of nuclear fusion has also entered a critical period of its development. How to let more people to understand nuclear fusion and join in its research is one of the effective means to accelerate the implementation of fusion. This paper proposes the first large model in the field of nuclear fusion, XiHeFusion, which is obtained through supervised fine-tuning based on the open-source large model Qwen2.5-14B. We have collected multi-source knowledge about nuclear fusion tasks to support the training of this model, including the common crawl, eBooks, arXiv, dissertation, etc. After the model has mastered the knowledge of the nuclear fusion field, we further used the chain of thought to enhance its logical reasoning ability, making XiHeFusion able to provide more accurate and logical answers. In addition, we propose a test questionnaire containing 180+ questions to assess the conversational ability of this science popularization large model. Extensive experimental results show that our nuclear fusion dialogue model, XiHeFusion, can perform well in answering science popularization knowledge. The pre-trained XiHeFusion model is released on https://github.com/Event-AHU/XiHeFusion.

Problem

• Develops XiHeFusion for nuclear fusion communication
• Enhances logical reasoning with chain of thought
• Assesses model with 180+ question test questionnaire

Innovation

• Supervised fine-tuning of Qwen2.5-14B
• Chain of thought for logical reasoning
• Multi-source knowledge training support