Startup Concept Generator With GPT-J

Generate concept for a startup based on keywords - GPT-J Completion

No time to walk through the tutorial? Test this prompt out here.

Use Case

A fun way to get started with building AI text applications can be making one that will generate startup ideas for you from examples of keywords and descriptions. The keyword is a word or two words describing the problem you aim to solve, while the description can include a sentence summary of the application and features of the application you intend to build.

If you don’t have an account, create an account on the Mantium Platform.

When you are ready to create your prompt, click AI Manager > Prompts > Add New Prompt, and fill out the following:

  • Name of Prompt: Startup Generator
  • Description: Generate concept for your startup based on keywords

Tags and Intelets can be left blank.

For deploying your prompt publicly at the end of this tutorial, you can add a default security policy configured by Mantium. Click Add Security Policies under Security Policies and drag Default Policies from All Policies to Selected Policies. Click Done to save - you will know the policy has been applied when its name is visible under Security Policies.

Provider Settings

  • Provider: Mantium
  • Endpoint: Completion
  • Model: GPT-J

Prompt Body

We have prepared a text snippet(see below) for you to paste into the Prompt Text field — it contains examples of Keywords and Descriptions for four startup concepts. You can edit these examples as you see fit, you can also add more examples.

Note that when pasting, be sure to add two lines after the last Description (See Image below). You can do this by pressing the Enter or return key twice. Language models are sensitive to all characters, including spaces.

Keyword: Event Planning
Description: A mobile application that helps with event planning. Its features include calculators for venue capacity, staffing, catering, staging, projection and dance floor.

Keyword: Delivery Service
Description: A web and mobile application that helps users who are too busy to shop, and offers deliver services. It features include product lists, order tracking, schedule delivery, payment option, customer feedback and dashboard.

Keyword: Children Learning
Description: A mobile application where kids can learn how to code. It features include learning management system with contents, quiz taker, fun games and parent dashboard.

Keyword: Internet
Description: A web application that helps users to connect with local businesses. It features include business directory, directory of local businesses, map with business information, job board and more.

Prompt Settings

  • Min Length: 15
  • Max New Tokens: 40
  • Temperature: 0.7
  • Top P: 1
  • Stop Sequences: \n

Max Tokens controls the length of an output response. The count should be adjusted for the general expected length of the input and output.

Temperature controls “creativity” - higher temperatures will produce more creative outputs but are less likely to adhere to structure. A lower temperature is advised for a prompt that requires a well-defined response, as the model will choose words with a higher probability of occurrence.

Top P is another way to control "creativity" using a different probability method. Setting Top P to .1 means that the model will only consider the words with top 10% probability masses as potential outputs.

Stop sequences are another method of controlling output - they allow you to define any text sequences that force the model to stop. If you don't have a stop sequence, the model might generate a stream of the requested response length, or stop in the middle of a sentence.
Using "\n" illustrates the new line between the examples. Note that at the end of the examples, we still have a new line, after which the model takes the Test Input in the following line and generates a completion for this input.

Test Prompt Text

Because these models are stochastic, the resulting output will vary even when using the same input, but we have included a text snippet to run as a test. Notice that after the : there is no space. (See Image below)

Keyword: Financial Technology 
Description:

You can either test this using the input field above, or deploy the prompt, and test using the deployed prompt input field. (See how to deploy below)

There is still a chance that using the same input multiple times may yield a different output. However, configuring the prompt with a low temperature and small value for Top P will help increase consistency of the output. We suggest clicking Test Run multiple times and tweaking the values of Temperature and Top P to test output consistency as needed. Once you are satisfied with the results, click Save.

One-Click Deploy

Mantium enables sharing your prompt using the One-Click Deploy feature, located in each prompt's drawer view. From your list of prompts, click on the Startup Generator prompt, click Deploy , and add the following configuration settings:

  • Name: Startup Generator
  • Description: Generate concept for your startup based on keywords
  • Author Name: Your Name
  • ✅ Add Input Field
  • Public
  • Live
  • Input Placeholder Text: Enter a Keyword for your startup, and Description

Prompt Styling

You can customize the single-page application rendered after deploying the prompt. To do this, you can click on the Show Preview button, and then select a preset theme or edit the theme colors based on your specification.

You can preview your styling with the Preview button or save your styling with the Save Styles button.

After deploying your prompt, you will have a unique URL that you can share for others to interact with it. To interact with the prompt, a user should input a startup keyword, and the Description keyword with : (See the test input example above)

Testing

To test out the prompt that we have configured, click this link!

Below is my deployed prompt. Notice that you can view the examples by clicking the Examples tab at the bottom left corner to view the input example.

Similar Use Cases

Another example of a concept generator use case is Game Concept Generator. You can find the tutorial here to learn how to build this text application.


Did this page help you?