Verdict
Type
Task
Reviewer
Review status
Robot
Day
-
Duration (seconds)
-
QA checks
Data collection
hours collected per day, all programs · green = accepted, red = not yet acceptedOperator leaderboard
hours collected and on the clock, per operator (PST)By day
bar length = hours collected, colors = verdict mixWhy episodes fail
verdict mix, check severity, then which checks fireVerdict mix
Checks fired
Top checks
Where gates trip, by operator
episodes each check fired on, per operator · stronger = moreTrends
day-over-day & week-over-weekRegrade admin
Rubric version:
loadingto