Just a screenshot of it being stupid while sounding smart
128M parameters 21B tokens training data kinda stupid, but honestly not that bad, it even got the markdown formatting right
