Research on Commercial Cryptography Large Language Model Based on Retrieval Augmented Generation and Low-Rank Adaptation Technology

As large language model technology improves by leaps and bounds, artificial intelligence has been initially applied in the field of the law and medical education. However, commercial passwords have not been widely used, resulting in the failure of information security of citizens and enterprises. This paper proposes a large language model of commercial cryptography based on search retrieval augmented generation and low-rank adaptation technology, aiming to improve the application of commercial passwords. The results show that the model is more effective than GPT4 in answering questions in the field of commercial cryptography.