EvaluatorGenerationInputs Class
Definition
Important
Some information relates to prerelease product that may be substantially modified before it’s released. Microsoft makes no warranties, express or implied, with respect to the information provided here.
Caller-supplied inputs for an evaluator generation job.
public class EvaluatorGenerationInputs : System.ClientModel.Primitives.IJsonModel<Azure.AI.Projects.EvaluatorGenerationInputs>, System.ClientModel.Primitives.IPersistableModel<Azure.AI.Projects.EvaluatorGenerationInputs>type EvaluatorGenerationInputs = class
interface IJsonModel<EvaluatorGenerationInputs>
interface IPersistableModel<EvaluatorGenerationInputs>Public Class EvaluatorGenerationInputs
Implements IJsonModel(Of EvaluatorGenerationInputs), IPersistableModel(Of EvaluatorGenerationInputs)- Inheritance
-
EvaluatorGenerationInputs
- Implements
Constructors
| Name | Description |
|---|---|
| EvaluatorGenerationInputs(String, IEnumerable<EvaluatorGenerationJobSource>, String, String) |
Initializes a new instance of EvaluatorGenerationInputs. |
Properties
| Name | Description |
|---|---|
| Category |
Category determines the rubric generation focus: 'quality' (default) produces quality-focused rubric criteria, 'safety' produces policy-derived safety rubric criteria. Both use the same rubric structure. Singular because quality and safety generation are mutually exclusive pipelines — the output EvaluatorVersion.categories is an array (e.g., ['agents', 'quality']). |
| EvaluatorName |
The evaluator name to create or update. If an evaluator with this name already exists, the service retrieves the latest version's criteria as context for improvement. |
| Model |
The LLM model to use for rubric generation (e.g., 'gpt-4o'). Required — users must provide their own model rather than relying on service-owned capacity. |
| Name |
Display name for this generation job. |
| Sources |
Source materials for generation — agent descriptions, prompts, traces, or datasets. Each entry is an |
