The AI models used on here treat the word "BREAK" just like…

The AI models used on here treat the word "BREAK" just like any other word in the prompt - it does not serve as a function to separate your prompt. I generally structure my prompts in a certain order. I sort of ask myself these questions and the answers are the prompt: "What do I want a see?" (the subject) "What do they look like?" (its description from broadest descriptors down to the more detailed aspects if I care) "What are they doing?" (the pose and/or action) "Where are they?" (location details) Lastly, I give the style, lighting, and camera angles if I care: "highly realistic, detailed skin, cinematic lighting, view from above." I also try to condense descriptors if I can - hair in particular. So instead of "long hair, black hair, wavy hair" I'd enter "long black wavy hair" - grouping them and following a pattern helps. If in the middle of describing the subject I tossed in something about the location then continued to describe the subject and later added more about the location, it seems to confuse the AI sometimes. Also using weights helps get the most important elements right. ((long black wavy hair: 1.3)) ((red neon lighting: 1.7)) will definitely ensure the AI captures those above all. If I'm still not getting what I want and don't want to mess around with the other settings too much, I just dial back on the prompt details to only enter the most important elements and slowly add details over time. I hope that helps!