The result of evaluate: a score, its range, and every answer.
evaluate
Readonly
The id of this run: its calls carry it as caller.evaluation in the call log.
caller.evaluation
The first metric's mean, from 0 to 1.
One line per metric: the mean, its 95% range, how many rows, how many failed.
Each row's value for a metric (the first by default); a failed row counts 0.
The result of
evaluate: a score, its range, and every answer.