Result from a judge evaluation containing score, reasoning, and metadata.

Hierarchy

  • LDJudgeResult

Properties

errorMessage?: string

Error message if evaluation failed

judgeConfigKey?: string

The key of the judge configuration that was used to generate this result

metricKey?: string

The metric key for this evaluation

reasoning?: string

Reasoning behind the provided score

sampled: boolean

Whether this evaluation was sampled (i.e. actually run). False when skipped by sampling.

score?: number

Score between 0.0 and 1.0 indicating the evaluation result

success: boolean

Whether the evaluation completed successfully

Generated using TypeDoc