I would like the ability to have power over prompts.

The AI's mislearning is too strong and does not follow the prompts.

It is very difficult to erase the mis-learning once it is created. Once unnecessary elements are generated, even if negative prompts are used or instructions are given not to draw 00, they will definitely be drawn.

Even if you enter "crying woman" or "smiling man," a "laughing woman" or "crying man" will be drawn. No matter how much I use {{act}}, {{must}}, {{{}}}, or negative prompts, I cannot correct the mislearning.

I wish you would stop reflecting other factors when you limit it to "it's a woman who is crying" in the first place.

From the above, I want the prompts to have more power than the AI.