RationaleResponse

data class RationaleResponse(val rationale: String, val confidence: <Error class: unknown class> = 0.7)

What the LLM sends back: the generated rationale prose and its self-reported confidence.

The confidence field defaults to 0.7 when the model omits it and is clamped to 0.0, 1.0 before being written into the RationaleArtifact.

Constructors

constructor(rationale: String, confidence: <Error class: unknown class> = 0.7)

val confidence: <Error class: unknown class> = 0.7