Medicine based on Symptoms: Improving Large Language Models to Answer Multiple Choice Questions

In recent years, Large Language Models (LLMs) have become known for answering multiple different types of questions. However, these models give the wrong answer many times, and that may be a problem in cases where mistakes may cost lives. In this paper, I present ways to help increase the accuracy of multiple models based on the MedQA dataset. I have trained models on textbooks, Q&As, and prompts to have a fuller understanding of what methods work the best. These AIs in the field of Biomedics have a prominent future and are currently understudied. If LLMs can learn how to accurately answer multiple choice questions, they will be able to help people from all over the world, and all from home. With the methods explained in the paper, the accuracy of the model has increased over 15 percent with the highest being a model 55 percent accuracy. Overall, the MedQA dataset presents great challenges to multiple models, and I hope to use the dataset to provide a deeper understanding of the methods that can be used to increase the accuracy of LLMs.