setupLLMAsAJudge

suspend fun <T> AIAgentGraphContextBase.setupLLMAsAJudge(task: String, llmModel: LLModel?, nodeInput: T): CriticResult<T>(source)

InternalAgentsApi method. Configures llmAsAJudge

Return

A result of type CriticResult containing the evaluation outcome, feedback generated by the LLM, and a reference to the original input.

Parameters

task

A description of the evaluation task that the LLM will perform. This task defines the scope and criteria for judging the input.

llmModel

The LLM model to use for the evaluation. If null, the default model will be used.

nodeInput

The input data to be evaluated by the LLM according to the specified task.