Skip to content
TheChatArchiveThe Chat Archive
gpt-4-1106-preview· OpenAIHugging Face Arena Human Preference 55K

improve my prompt so my ai image generator actually shows what i want it to show…

From a Chatbot Arena battleThe same person sent this prompt to two anonymous models at once. Compare it with gpt-3.5-turbo-0613’s answer.
  1. User

    improve my prompt so my ai image generator actually shows what i want it to show "Melissa McCarthy with pet harper seal sitting on her dirty couch with a very jealous looking cat sitting overtop giving them an angry stare she has a remote in her hand and she cleaned her house so there are only a few pieces of trash and cardboard boxes on the floor now she is trying to be the crazy seal lady instead of crazy cat lady but she still looks crazy"

  2. Assistant

    To improve the clarity and specificity of your prompt for an AI image generator, consider breaking down the scene into clear, visual descriptions that focus on the composition, expressions, and key elements you want to capture. Here's a refined version of your prompt:

    "Create an image featuring actress Melissa McCarthy sitting on a well-worn, somewhat dirty couch. She is joined by her adorable pet harbor seal, which is lounging contentedly beside her. In contrast, there is a very jealous-looking cat positioned above them on the backrest of the couch, glaring down with an angry stare. Melissa holds a TV remote in one hand, suggesting she's ready to watch her favorite show. Despite her efforts to tidy up, there are still a few stray pieces of trash and cardboard boxes scattered on the floor around the couch, hinting that she's in the process of cleaning. The overall scene should exude a quirky charm, with Melissa embracing her new identity as the 'crazy seal lady,' yet still retaining an air of her previous 'crazy cat lady' persona."

    This revised prompt provides clear guidance on the subjects (Melissa, the seal, the cat), their expressions and interactions, as well as the setting and mood of the scene. It should help the AI generate an image that better matches your vision.

Responses were produced in different, uncontrolled contexts (system prompts, settings, prior turns). This is not a benchmark.