Introduction Latent Space in AI is the compressed, lower-dimensional representation of data used in AI to capture essential features and patterns. Where similar points cluster together closely. AI uses this space to make meaningful connections and generate outputs based on the patterns it has processed. I’ve made an interesting testable observation; the tone of input can influence the depth, elaboration and style of an AI’s response. We have all heard of prompt engineering, this focuses heavily on the precision and descriptiveness of a prompt. But tone is often overlooked. So, how does the tone of a prompt affect AI responses, and what does this reveal about latent space utilisation? Method/ Experiment I conducted a small and replicable study which you can reproduce with any model. I used ...
No Comments 
