Enhance your Triplo AI experience with the new "Training" feature. By creating a MIND, you can provide Triplo AI with a wealth of knowledge that it will use to craft more accurate and relevant responses for you.
What is a MIND?
  • A MIND is a collection of up to 200,000 tokens of text (roughly equivalent to two romance novels) that you can build and Triplo AI will feed the AI model to serve as context for your prompts hence enhancing the quality of content generated.
  • Think of a MIND as a personalized knowledge base accessible by Triplo AI to reference AI when you prompt.
How to Use Training
  • Create a MIND: Click the brain-shaped "MIND icon" on the main Triplo AI interface and then the green "Create New Mind" button. Give your MIND a title and paste your training data into the "Training Data" field.
  • Add Content: To add more to a MIND, click the PLUS sign next to its title. You'll see how many tokens you can still add below the data field.
  • Activate a MIND: Once your MIND is ready, activate it by selecting it from the Training screen or directly via the MIND icon.
Costs and Considerations
  • Training a MIND is cost-effective, with each token being five times less expensive than regular prompting.
  • When using your own API keys, the cost is minimal (almost irrelevant).
  • When activated a training generally adds between 500 and 1000 tokens to your prompt as context.
  • Keep in mind the 8,000-token training window: if you train less, you're only charged for the tokens used, but the MIND's limit decreases by the full 8,000.
Best Practices
  • Only activate MINDs when necessary, as they are tailored to feed prompts related to the content you trained (specific content).
  • Currently, MINDs accept only textual input, but we're expanding to include direct uploads from files and URLs soon.
  • The Training feature is currently available on the Desktop version of Triplo AI, with mobile support on the way.
Remember, a MIND is not for processing (recognizing styles or learning processes, or anything related to the model structure); it's a powerful context provider for generating content that matters to you.

Was this article helpful?