View Source ExOpenAI.Components.Eval (ex_openai.ex v2.0.0-beta2)

An Eval object with a data source config and testing criteria. An Eval represents a task to be done for your LLM integration. Like:

Improve the quality of my chatbot
See how well my chatbot handles customer support
Check if o4-mini is better at my usecase than gpt-4o

Fields

:created_at - required - integer()
The Unix timestamp (in seconds) for when the eval was created.
:data_source_config - required - map()
Configuration of data sources used in runs of the evaluation.
:id - required - String.t()
Unique identifier for the evaluation.
:metadata - required - ExOpenAI.Components.Metadata.t()
:name - required - String.t()
The name of the evaluation.
:object - required - :eval
The object type.
Allowed values: "eval"
Default: "eval"
:testing_criteria - required - [ExOpenAI.Components.EvalGraderLabelModel.t() | ExOpenAI.Components.EvalGraderStringCheck.t() | ExOpenAI.Components.EvalGraderTextSimilarity.t() | ExOpenAI.Components.EvalGraderPython.t() | ExOpenAI.Components.EvalGraderScoreModel.t()]
A list of testing criteria.
Default: "eval"

Summary

Types

t()

Types

t()

@type t() :: %ExOpenAI.Components.Eval{
  created_at: integer(),
  data_source_config: map(),
  id: String.t(),
  metadata: ExOpenAI.Components.Metadata.t(),
  name: String.t(),
  object: :eval,
  testing_criteria: [
    (((ExOpenAI.Components.EvalGraderLabelModel.t()
       | ExOpenAI.Components.EvalGraderStringCheck.t())
      | ExOpenAI.Components.EvalGraderTextSimilarity.t())
     | ExOpenAI.Components.EvalGraderPython.t())
    | ExOpenAI.Components.EvalGraderScoreModel.t()
  ]
}