No less important are negative prompts - another example of economy and efficiency. You don't need to fill up with negative tags, especially if you use Embeddings (Spells). Spells are LoRA in essence, that is, Low-Rank Adaptation of Large Language Models, these LoRA are pre-trained model weights and inject trainable layers (rank-decomposition matrices) in each transformer block, that is, in the case of Negative Prompts, they already include a sequence of information.

In this simple image, I only used 1 negative: easynegative

These embedding learn what disgusting compositions and color patterns are, including faulty human anatomy, offensive color schemes, upside-down spatial structures, and more. Placing it in the negative can go a long way to avoiding these things.

So it's obvious that the negative prompt also influences the final result. So instead of using a bunch of noise prompts like: ((worst quality, low quality:2),monochrome,overexposure,watermark,text,bad anatomy,bad hand,((extra hands)),inverted hand, disembodied limb,oversized head, extra body, extra navel etc...

Just use one or two Spells, I recommend easynegative + negative Hand-neg. This leaves room to fix problems, for example. Is your image coming with a signature? Include the negative prompt: (signature)

My recommendation in most of cases would be: polydactyly, negative_hand-neg, ng_deepnegative_v1_75t, (watermark, logo, signature, extra limbs)

Hardly you will need more than that, but, each case is a single case.

Due to the 2k character limits per post, to see all 10 parts of the guide about prompts, go to my profile: https://yodayo.com/posts/01a5e690-eabf-45cc-8309-4d0f931b7178/