Skip to main content

​
What these models are

LearnTerms generation models are product-facing modes for AI-assisted question creation. They are part of the LearnTerms generation layer and are meant to hide raw model complexity behind subject-aware choices that make sense to contributors.

​
Available models

​
How to choose

Pick the model that matches the course or material you are working with.
  • Use swift-general when the material spans multiple domains.
  • Use swift-optometry when the content is specifically vision or ocular focused.
  • Use swift-pharmacy when the source material centers on therapeutics, mechanisms, and medication use.

​
What the generation layer is trying to do

The LearnTerms layer is not meant to act like a generic chatbot. It is designed to produce study questions that better match course context, academic domain, and exam style. That means the model choice should stay close to the subject matter, not just user preference.

​
Practical guidance

  • Start with the model that most closely matches the module topic.
  • Use stronger source chunks before you try to tune prompts.
  • Treat model choice as a relevance decision, not a creativity setting.