Skip to main content

Interface: EvaluateOptions

evaluation.EvaluateOptions

Properties

data

data: DataT

The dataset to evaluate on. Can be a dataset name, a list of examples, or a generator of examples.

Defined in

evaluation/_runner.ts:76


client

Optional client: Client

The LangSmith client to use.

Default

undefined

Defined in

evaluation/_runner.ts:106


evaluators

Optional evaluators: EvaluatorT[]

A list of evaluators to run on each example.

Default

undefined

Defined in

evaluation/_runner.ts:81


experimentPrefix

Optional experimentPrefix: string

A prefix to provide for your experiment name.

Default

undefined

Defined in

evaluation/_runner.ts:96


maxConcurrency

Optional maxConcurrency: number

The maximum number of concurrent evaluations to run.

Default

undefined

Defined in

evaluation/_runner.ts:101


metadata

Optional metadata: KVMap

Metadata to attach to the experiment.

Default

undefined

Defined in

evaluation/_runner.ts:91


summaryEvaluators

Optional summaryEvaluators: SummaryEvaluatorT[]

A list of summary evaluators to run on the entire dataset.

Default

undefined

Defined in

evaluation/_runner.ts:86


Was this page helpful?