The realm of artificial intelligence opens a constant flux of novel models. These models, sometimes exposed prematurely, provide a unique opportunity for researchers and enthusiasts to scrutinize their inner workings. This article delves into the practice of dissecting leaked models, proposing a