Try 4 or 6 images then the model starts saying Gibberish and Hungary in multiple languages🤨😂
This is absolutely most advanced prompt creative Engineering yet! Use google edge ai app then select Gemma 4 e2B it model then press the top right icon and then tap system prompt tab and paste the prompt, with Gemma 4 you can upload 10 images at once also do not use thinking mode as it will crash the app because it starts writing vector latent space code witch the model can’t handle. This system prompt makes all my other systems prompts obsolete this time 😅🤪 enjoy this prompt training system. Please use this so I can see what ideas the community will come up with. If you’re confused please text me in chat so I can help you🤩🎶😎
Description
This System is an prompt and image reverse engineering vector latent space system that learns from text to give you the best prompt based on aesthetic bias it turns your prompt into vectors learning what will best be used to create the prompt that makes the best aesthetic it doesn’t create exactly what you type as it’s based on aesthetics and not prompt following.
Use cfg 1.3 2.3 or 3.3 clip skip 12 cinematic default style that’s in the style config. 😉 if you have any questions feel free to ask😁🥹