Overview of the development and applications of large language models
WANG Wenqi
GUO Mengfan
YANG Duxiang
ZHANG Jingyuan
HONG Feiyang
Abstract：With the rapid development of AI,large-scale models have emerged as a driving force be-hind scientific and technological progress as well as industrial transformation,attracting widespread attention from both academia and industry.As large-scale pre-trained models,large language models(LLMs)based on deep learning technology possess an enormous number of parameters and exhibit powerful learning and generalization abilities,enabling them to process and generate various types of data.Based on a review of the development of LLMs,this paper discusses the architecture design cen-tered around Transformer,analyzes current training and optimization techniques such as supervised fine-tuning and reinforcement learning alignment,and explores the practical applications and develop-ment of LLMs.Taking the DeepSeek series as an example,this paper examines the innovations and applications of domestic LLMs.Additionally,the challenges faced in real-world applications,such as model hallucinations,efficiency bottlenecks,and value alignment bias,are briefly discussed,and fu-ture development trends of LLMs are outlined.
Keywords：large language modelsdeep learningTransformerDeepSeekmodel hallucinations
Publication Date：2025-04-25
Online Publishing Date：2025-08-15(First online date of this platform, not the publication date of the document)
Pages：8( 1-8 )