REVIEW 9 cited by
Understanding the Impact of Negative Prompts: When and How Do They Take Effect?
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
read the original abstract
The concept of negative prompts, emerging from conditional generation models like Stable Diffusion, allows users to specify what to exclude from the generated images.%, demonstrating significant practical efficacy. Despite the widespread use of negative prompts, their intrinsic mechanisms remain largely unexplored. This paper presents the first comprehensive study to uncover how and when negative prompts take effect. Our extensive empirical analysis identifies two primary behaviors of negative prompts. Delayed Effect: The impact of negative prompts is observed after positive prompts render corresponding content. Deletion Through Neutralization: Negative prompts delete concepts from the generated image through a mutual cancellation effect in latent space with positive prompts. These insights reveal significant potential real-world applications; for example, we demonstrate that negative prompts can facilitate object inpainting with minimal alterations to the background via a simple adaptive algorithm. We believe our findings will offer valuable insights for the community in capitalizing on the potential of negative prompts.
Forward citations
Cited by 9 Pith papers
-
ReNeg: Learning Negative Embedding with Reward Guidance
ReNeg optimizes a negative text embedding with reward feedback and classifier-free guidance in the training loop, improving image-video generation quality over null-text and handcrafted negative prompts.
-
SimVS: Simulating World Inconsistencies for Robust View Synthesis
Video diffusion models simulate world inconsistencies, and a harmonization network trained on the simulated data reconciles sparse inconsistent multi-view images into consistent 3D scenes.
-
Safeguarding Text-to-Image Generation via Inference-Time Prompt-Noise Optimization
Prompt-Noise Optimization jointly tunes the prompt embedding and diffusion noise at inference time to suppress unsafe images while keeping outputs close to the prompt.
-
Negative Token Merging: Image-based Adversarial Feature Guidance
NegToMe pushes each generated image token away from its closest matching token in a reference image during reverse diffusion, improving diversity and reducing copyright similarity without retraining.
-
ContrastiveCFG: Guiding Diffusion Sampling by Contrasting Positive and Negative Concepts
A new guidance reweighting, derived from a contrastive loss, makes negative prompting in diffusion models remove unwanted concepts with less quality loss than standard negated CFG.
-
Classifier-Free Guidance inside the Attraction Basin May Cause Memorization
Delaying classifier-free guidance until a 'transition point' in the denoising process reduces verbatim memorization in diffusion models, with a new 'opposite guidance' variant to escape memorization basins sooner.
-
CountGD++: Generalized Prompting for Open-World Counting
Counting with negative prompts, auto-generated exemplars, and external or synthetic example images improves accuracy across seven open-world counting benchmarks.
-
$\textit{Revelio}$: Interpreting and leveraging semantic information in diffusion models
Diffusion model features, when decoded with k-sparse autoencoders, reveal interpretable visual concepts, and a lightweight classifier on the best layer (up_ft1 at t=25) beats prior diffusion-based classifiers on fine-...
-
Stylecodes: Encoding Stylistic Information For Image Generation
An open-source method that compresses image style into a 20-character base64 code and uses it to condition Stable Diffusion image generation, with only qualitative evidence of quality.
Discussion (0). Continue with ORCID to comment.