The growing interest in large language models (LLMs) has led to exploring their potential in various domains, including medical research. This work focuses on using Llama-3 70B, an open-source LLM, to automate insights from neurodegenerative disease research papers. By applying zero-shot inference for hypothesis and result extraction, we aim to analyze multiple papers and find correlations in their methodologies. Our research seeks to enhance the understanding of complex findings, with an emphasis on accuracy and reliability. Currently, we are in the early stages, using a small open-source model to test a pipeline that shows promising results. We are evaluating a hand-picked set of highly cited papers from PubMed, published in the last four years, through a qualitative analysis comparing the model’s output to our interpretations.