Hiding an LLM configuration from Prompt Builder model selection
Key Takeaway
Hiding an LLM configuration only controls what users can select in Prompt Builder. Prompt templates that already use the hidden model keep working.
An LLM configuration is the model setup a prompt template uses to generate output in Prompt Builder. Hiding a configuration removes it from the models users can pick in Prompt Builder. Hiding doesn't turn the model off, so it's a safe way to steer users' future choices without breaking prompt templates that are live.
When to hide an LLM configuration
Hide an LLM configuration when you want users in Prompt Builder to switch their prompt templates to a different model. The different model could be a newer LLM version. It could also be a different LLM, possibly from a different provider.
You can also hide a new LLM to suspend access to it until you finish testing it. Hiding can be undone, because you can later show the hidden configuration so that it's available for selection again.
For example, imagine you have just set up a configuration for a new LLM, but you haven't finished testing it. You hide the configuration, so users in Prompt Builder can't select it for their prompt templates while you test. When testing is done, you show the configuration again, and users can select it.