This is one of the twenty-five questions a carrier reviewer sends a vendor, in the words an examiner uses. Below it: the answers that make a reviewer stop on this item, derived by running Vaulith's rules engine on an estate that gets everything wrong, and the free questions that reach it.
The training corpus is described in engineering documentation but has never been approved as a record. Because the model is consumer-impacting, it is the one an examiner is most likely to name, and the gap sits on exactly that model.
To fix: Produce a provenance record covering sources, date range, volume, de-identification step, and approval.
1 response depend on an artifact that does not exist. An answer with nothing behind it is an assertion, and a reviewer under time pressure treats an assertion as an open question.
To fix: Produce the artifact, or withdraw the claims that rest on it.
The card was approved 2025-01-15 and the model was retrained 2026-06-01, 17 months later. The content may be strong; the date is the problem. A reviewer cannot tell whether the document describes what is running now, and that question costs them four seconds to ask and costs you a week to answer once the review is already running.
To fix: Compare the approved card against the model that is running today, write down what changed, and have the new version approved.
A standard item carries weight 2. A pass earns 1.00 of it, a questioned answer 0.60, an escalated one 0.20 and a stop 0.00. An unanswered item earns nothing and still takes the citation penalty, so a profile cannot be improved by leaving an inconvenient question blank. The whole arithmetic is shown on the homepage and in every assessment.