LLMEvaluator#

classDiagram EvaluatorBase <|-- LLMEvaluator
class council.evaluators.LLMEvaluator(llm: LLMBase)[source]#

Bases: EvaluatorBase

Evaluator using an LLM to evaluate chain responses.

__init__(llm: LLMBase)[source]#

Build a new LLMEvaluator.

Parameters:

llm – model to use for the evaluation.

execute(context: AgentContext, budget: Budget) → List[ScoredChatMessage][source]#

Executes the evaluator on the agent’s context within the given budget.

Parameters:

  • context (AgentContext) – The context for executing the evaluator.

  • budget (Budget) – The budget for evaluator execution.

Returns:

A list of scored agent messages resulting from the evaluation.

Return type:

List[ScoredChatMessage]

Raises:

None –