init(instructions:evaluationTarget:reference:)
Creates a model-as-judge prompt configuration.
Declaration
init(instructions: String = ModelJudgePrompt.defaultInstructions, evaluationTarget: (@Sendable (Input.ExpectedValue) -> String)? = nil, reference: (nonisolated(nonsending) @Sendable (Input, Input.ExpectedValue) async throws -> [String : String])? = nil)Parameters
- instructions:
System instructions for the model-as-judge. Defaults to a general-purpose evaluator prompt.
- evaluationTarget:
Optional closure to convert the response to a string. When
nil, the response is JSON-serialized. - reference:
Optional closure returning labeled reference data to include in the judge prompt.
Discussion
let prompt = ModelJudgePrompt<ModelSample<String>>(
instructions: "You are a domain expert."
)