resolve_criteria#
- langchain.evaluation.scoring.eval_chain.resolve_criteria(criteria: Mapping[str, str] | Criteria | ConstitutionalPrinciple | str | List[Mapping[str, str] | Criteria | ConstitutionalPrinciple] | None) dict [source]#
Resolve the criteria for the pairwise evaluator.
- Parameters:
criteria (Union[CRITERIA_TYPE, str], optional) – The criteria to use.
- Returns:
The resolved criteria.
- Return type:
dict