Enhancing Antibiotic Question Answering with Large Language Models
Zewei Quan, Fenglian Yuan, Jingfei Fu, Wanwen Wu, Xiaobao Zhu · 2024
Abstract. This article explores LoRA and P-tuning v2 parameter fine-tuning techniques improve the ability of the Large Language Models (LLMs) in the antibiotics domain. Through specialized fine-tuning, a model named ANTI-CHATGLM was developed, which markedly enhanced the precision and reliability of responses to antibiotic-related questions. The experimental outcomes demonstrate that the proposed model surpasses the baseline model in metrics like BLEU and ROUGE, and had lower BARTscore. Comparative experiments and ablation analysis confirmed the model's effectiveness and application potential in antibiotic medical consultation services.