Arize Phoenix TS
    Preparing search index...
    • A function that leverages an llm to perform a classification

      When given an AI SDK evaluation model (e.g. TypeSafe's Jev) the classification is routed through experimental_evaluate as a single choice question. Evaluation models cannot generate text, so no explanation is returned in that case. telemetry, schemaName and schemaDescription are ignored on that path: experimental_evaluate does not emit spans yet.

      Parameters

      Returns Promise<ClassificationResult>