Form eval, side by side

Pick any two runs, A and B. Behaviours are grouped by bucket; open one to see why each run passes or fails, its graders and the scenarios they checked, down to each reply and the full conversation. Grey striped fails are expected for setups where Jev and code own goals and closing.

All behaviours at a glance

left half is A, right half is B passed failed Click a tile to open that behaviour, or a bucket name to open the bucket
Bucket› Behaviour› Grader› Scenario Global checks every scenario in the run Local checks only its behaviour's own scenarios passed failed check did not apply turn lost, not graded

Loading reports…

The team's original reports · the unchanged report.html for A and B, side by side · optional

Opening a behaviour above also opens it here.

A

Loading the full report (about 27 MB)…

B

Loading the full report (about 27 MB)…