The idea of stable diffusion has acquired a lot of momentum in the dynamic field of artificial intelligence due to its potential to boost language model performance in a variety of contexts. The use of negative prompts in stable diffusion is one particular area that has caught the interest of scholars. These negative prompts have become a potent weapon in 2023, enabling more precise and refined outputs based on the environment. In this article, we explore, “demystifying negative prompts in Stable Diffusion”, examining their importance, describing how they work, and offering illustrated instances.

Despite being a sophisticated AI picture generator, Stable Diffusion occasionally generates images that are hazy, deformed, or otherwise unrepresentative of your original expectations. Negative cues enter the picture at this point. Instructions on the Stable Diffusion Negative Prompts List inform the algorithm what you don’t want to see in an image.

You might, however, occasionally be dissatisfied with the results, which will keep you from acquiring the desired image. This is when negative cues are useful. This post will explain how to apply negative suggestions in stable diffusion, how they might enhance your photographs, and which negative prompts are most appropriate for certain kinds of images.

What is Stable Diffusion ?

Stable Diffusion is a deep learning model which makes use of diffusion techniques to create excellent artwork from input photos. Simply put, when you ask Stable Diffusion a question, the model has been taught to produce a plausible illustration of the object you describe.

By handling complicated and abstract text descriptions, it significantly outperforms earlier text-image generators. It accomplishes this by using a novel technique known as stable training, which enables the model to produce excellent visuals that are consistent with the textual input.

A variety of artistic mediums, such as photorealistic portraits, landscapes, and abstract art, can be created using stable diffusion. The algorithm has been utilised in a variety of projects, including making digital art, making video games, and producing photographs for scientific research.

Demystifying Negative Prompts in Stable Diffusion

The concept of “stable diffusion” emphasises the controlled and progressive introduction of various stimulus types during the model’s training. The model improves its ability to manage complex situations and produce nuanced responses by coming across a range of cues, including those that test its comprehension.

Overall, stable diffusion tries to improve language models by simulating a wide variety of inputs, allowing them to come up with outputs that are not only accurate in fact but also take into account varied viewpoints and eliminate biases.

Demystifying Negative Prompts in Stable Diffusion — What Are They ?

A negative prompt is a technique for using stable diffusion that enables the user to define what he does not want to see without providing any further input. This option instructs the Stable Diffusion model which should not be included in the image that is generated.

By serving as a high-dimension anchor that the process strays away from, negative prompting has an impact on how ideas are generated. It can also be utilised to enhance the output of the image by defining amorphous terms like “blurry” and “pixelated.”

By functioning as a high-dimensional anchor that the process strays from, negative prompting has an impact on how ideas are generated. This makes it possible to more accurately regulate the output image. Users can create distinctive graphics with greater detail and precision by employing negative cues.

For instance, even when you supply Stable Diffusion with settings like “Don’t add duplicates,” it could still produce copies even though you may have made a portrait. Because it comprehends negative cues more well than human language, this is.

Instead of reproducing the same question in such situation, you may offer a derogatory prompt like “duplicate” or “poorly Rendered face.”

Without a negative prompt, you will receive what is listed below:

What you will obtain if you include a derogatory prompt like “Duplicate” is as follows:

The Role of Negative Prompts

In order for steady diffusion to work, negative stimuli are a crucial component. These questions are intended to give the model examples of information that is false, deceptive, or erroneous. There are various benefits to including negative prompts in the training dataset.

1. Contextual Understanding

Negative prompts put the model’s capacity for precise context perception to the test. It forces the AI to consider several facets of the context before producing a response by exposing the model to false or contradictory information.

2. The reduction of bias and fairness

By using negative cues, the model is able to recognise and correct biases that exist in its training set. The model gains the ability to produce more fair and balanced responses as it is exposed to biassed stimuli.

3. Generating Counterarguments

Negative prompts allow the model to produce counterarguments or alternate viewpoints in instances when a response necessitates taking into account numerous perspectives. This results in richer and more thorough outputs.

How Do I Enter Negative Prompts?

Before creating an image, enter the prompts from Stable Diffusion Negative Prompts List 2.1 in the second text field. In the image below, the second text box is identified by a red circle. The primary question, which outlines what you want in your image, should be entered in the first text field.

The second text box is for the negative prompt, which instructs the AI what you don’t want in your image. After answering both questions, select Generate image to begin the image production process. With Stable Diffusion Negative Prompts List 2.1, negative prompts are an effective approach to customise your image output.

List of negative prompts for better images

If you wish to make any type of image using Stable Diffusion or another image generator, you can apply these general instructions to produce higher-quality photos. 

  • amputee
  • autograph
  • bad anatomy
  • bad illustration
  • bad proportions
  • beyond the borders
  • blank background
  • blurry
  • body out of frame
  • boring background
  • branding
  • cropped
  • cut off
  • deformed
  • disfigured
  • dismembered
  • disproportioned
  • distorted
  • draft
  • duplicate
  • duplicated features
  • extra arms
  • extra fingers
  • extra hands
  • extra legs
  • extra limbs
  • fault
  • flaw
  • fused fingers
  • grains
  • grainy
  • gross proportions
  • hazy
  • identifying mark
  • improper scale
  • incorrect physiology
  • incorrect ratio
  • indistinct
  • kitsch
  • logo
  • long neck
  • low quality
  • low resolution
  • macabre
  • malformed
  • mark
  • misshapen
  • missing arms
  • missing fingers
  • missing hands
  • missing legs
  • mistake
  • morbid
  • mutated hands
  • mutation
  • mutilated 
  • off-screen
  • out of frame
  • out of frame
  • outside the picture
  • pixelated
  • poorly drawn face
  • poorly drawn feet
  • poorly drawn hands
  • printed words
  • render
  • repellent
  • replicate
  • reproduce
  • revolting dimensions
  • script
  • shortened
  • sign
  • signature
  • split image
  • squint
  • storyboard
  • text
  • tiling
  • trimmed
  • ugly
  • unfocused
  • unattractive
  • unnatural pose
  • unreal engine
  • unsightly
  • watermark
  • written language

Best Negative Prompts for Different Situations

You could wish to include negative prompts that are more appropriate for the kind of art you’re about to produce when you’re drawing pictures in a certain style or setting. If you know how to add them to the generator you select, you may also use these keywords to generate images using other AI image tools.

Regarding Landscapes

You can use the following terms as negative prompts while making photos of a landscape, a work of natural beauty, or a picturesque view on Stable Diffusion:

  1. Blurry
  2. Boring
  3. Close-up
  4. Dark (optional)
  5. Details are low
  6. Distorted details
  7. Eerie
  8. Foggy (optional)
  9. Gloomy (optional)
  10. Grains
  11. Grainy
  12. Grayscale (optional)
  13. Homogenous
  14. Low contrast
  15. Low quality
  16. Lowres
  17. Macro
  18. Monochrome (optional)
  19. Multiple angles
  20. Multiple views

For Street Views and Cityscapes

You can use the following terms as negative prompts while creating photos of cities, streets, buildings, and monuments:

  1. Animals (optional)
  2. Asymmetrical buildings
  3. Blurry
  4. Cars (optional)
  5. Close-up
  6. Creepy
  7. Deformed structures
  8. Grainy
  9. Jpeg artifacts
  10. Low contrast
  11. Low quality
  12. Lowres
  13. Macro
  14. Multiple angles
  15. Multiple views
  16. Overexposed
  17. Oversaturated
  18. People (optional)
  19. Pets (optional)
  20. Plain background
  21. Scary
  22. Solid background
  23. Surreal
  24. Underexposed
  25. Unreal architecture

For Photos That Include Portraits of People and Animals

On Stable Diffusion, you can make photos of people and animals that look realistic as long as you use the following terms as negative prompts:

  1. 3D
  2. Absent limbs
  3. Additional appendages
  4. Additional digits
  5. Additional limbs
  6. Altered appendages
  7. Amputee
  8. Asymmetric
  9. Asymmetric ears
  10. Bad anatomy
  11. Bad ears
  12. Bad eyes
  13. Bad face
  14. Bad proportions
  15. Beard (optional)
  16. Broken finger
  17. Broken hand
  18. Broken leg
  19. Broken wrist
  20. Cartoon
  21. Childish (optional)
  22. Cloned face
  23. Cloned head
  24. Collapsed eyeshadow
  25. Combined appendages

Example #1

Without using negative prompts, one can produce a realistic image of a young man standing next to a red Ferrari.

Example #2

Add negative prompts like “ugly, tiling, poorly drawn hands, poorly drawn feet, poorly drawn face, out of frame, extra limbs,” to this result to improve it.

You can clearly observe the significant difference in these outcomes when using negative prompts or not.

As a result, it is both an optional and crucial factor in creating high-quality art.

How to make beautiful stable diffusion images with negative Prompts:

Best negative Prompts for stable diffusion:


Integration of negative stimuli has become a game-changer in the field of stable diffusion. They play a key role in creating contextually correct, unbiased, and thorough responses by challenging, improving, and enhancing language model outputs. The examples given in this article serve as a reminder of the impact that unfavourable cues will have on AI-generated content in 2023 and beyond. The ability to use negative prompts inside stable diffusion will surely open the way for more complex and ethical AI applications as this field of study continues to advance.



Are Negative Prompts Applicable in All Stable Diffusion Versions?

Every Stable Diffusion version utilises negative prompts since they are essential to producing beautiful artwork.

It’s more likely that the newest Stable Diffusion 2.0 will persuade you to employ Negative Prompts.

Does Using Only Positive Prompts Help Produce More Accurate Work?

Yes, it is possible, however the outcome and processing time may vary from image to image.

While some prompts are well-directed to generate realistic artwork, others call for the use of Negative Prompts, seeds, numerous samplings, and stages to obtain an accurate outcome.

Rohan Pradhan

Similar Posts

Leave a Reply

Your email address will not be published. Required fields are marked *