Over the past couple of months, I’ve been on a journey to understand protein language models - what exactly are they learning and how do they work?