Module 4 · ~30 min · draft
Deciding what to trust
- V1 Check a model's factual claims against a source and mark which held up.
- V3 Decide when to trust, verify, or reject a model output based on stakes and reversibility.
The task
Take a model output from the previous module that contains factual claims. Check three claims against real sources. Then classify the whole output: safe to use, use after fixes, or reject, and say why, based on what is at stake and how reversible a mistake would be.
What to submit
A claim-check table (claim, source, held or failed) and your trust decision with reasoning.
How it's graded
- Sources are real and cited.
- At least one claim was actually checked, not "seems right."
- The trust decision references stakes or reversibility.