Large Language Model？

As a significant breakthrough in the realm of generative artificial intelligence, large language models (LLMs) are increasingly permeating educational landscapes and revolutionizing teaching and learning. LLMs possess robust semantic comprehension, text generation and image generation capabilities that surpass traditional natural language processing methods. The application based on LLMs not only provides students with abundant learning resources but also offers diverse learning scaffolds to guide them in problem-solving. Additionally, LLMs can be utilized for automated teaching diagnosis, mitigating teachers' workload while enabling them to grasp learning situations efficiently, adjust teaching strategies, and devise more effective teaching evaluation methods.

Currently, a wide range of LLMs-based applications is emerging, including language translation, oral training, essay correction, programming tutoring, automatic problem-solving, automatic problem-solving, image generation. Nevertheless, the technology of LLMs has not yet matured, and some limitations and errors may occur, which adds barriers to the ground application of LLMs in education. Meanwhile, students' over-reliance on such tools may also lead to negative impacts. In addition, there is still a lack of evidence on how to properly and effectively utilize various types of assistants based on LLMs, and whether such tools can truly enhance learning. Therefore, further empirical research is urgently needed to address these issues.