Research on logical reasoning and question answering of large language models based on LoRA fine-tuning and prompt learning

With the continuous advancement of natural language processing technology, logical reasoning question and answer technology has received more and more attention. In this paper, we propose a systematic framework for evaluating and enhancing the logical reasoning ability of large language models. Firstly, we use the GLM-4-9B model as a base model, and the LogiQA dataset, which has a more complex dataset question form and very intrusive question options, as a benchmark dataset, to improve the model's reading comprehension ability through extensive training. At the same time, LoRA fine-tuning technique was used to compare the overall performance of the model before and after LoRA fine-tuning, and a model with stronger logical reasoning ability was obtained. Finally, a multi-faceted logical reasoning problem prompt function was designed to tap the logical reasoning potential of the large language model. The results show that the method of integrating LORA fine-tuning and cue learning can effectively improve the accuracy and performance of the model on logical reasoning tasks, which is an important reference value for future follow-up research on generative large language models