Page MenuHomePhabricator

[Testing] Manual test set for credibility signals v1.1
Closed, ResolvedPublic


Goal: To get a high level understanding of the effectiveness of the credibility signals v1.1 work.

Task: In a spreadsheet -- Ryan, Stephan, and Francisco will walk through 50 events and the "correctness" of the values provided in the realtime streaming API for the credibility signals v1.1 work.

They will record incorrect (red) and potentially incorrect (yellow) labelled in the correlated box in the spreadsheet for the event. If the result was a notable anomaly, add a note to what was seen.

We'll start with English (Ryan), Ukrainian (Stephan), and Spanish (Francisco).


Related Objects