Distribution statistics for a single metric across all evaluated samples.
Computed from the per-sample scores and available on EvaluationResult.stats.
Useful for understanding score variance, identifying outlier samples, and
reporting confidence in aggregate scores.
Distribution statistics for a single metric across all evaluated samples.
Computed from the per-sample scores and available on
EvaluationResult.stats. Useful for understanding score variance, identifying outlier samples, and reporting confidence in aggregate scores.