intermediate
Checks around the model that block harmful input, leaks and unsafe actions.
Bouncers at the doorBouncers at the door Why?: they stop bad requests in and private data out.
Here: What AI is (and isn't) · Next: How LLMs work
The main tools by job. Tap one to see what it does, then learn it here.
Eval frameworks
Tracing and experiments
Methods
Safety and guardrails