You can evaluate the correctness of various flows using custom Laminar pipelines. The only requirements are a dataset to run the evaluations against and a pipeline that produces a numeric output.

There are two possible setups for evaluations:

  1. Running Laminar pipeline and evaluating its results with another pipeline Learn more
  2. Running an evaluator pipeline on a dataset. Learn more

Status and results of the evaluation

Possible statuses of evaluations:

  • Started
  • Finished

Individual datapoint statuses “Success” / “Failed” are stored at datapoint level.