Surrogate Models

Fiddler’s explainability features require a model on the backend that can generate explanations for you.

📘 If you don’t want to or cannot upload your actual model file, Surrogate Models serve as a way for Fiddler to generate approximate explanations.

A surrogate model is built automatically when you call add_surrogate on an existing model that has a baseline dataset defined. You just need to provide a few key details on how your model operates during onboarding.

Surrogate Model prerequisites:

An onboarded model with:
- A defined model task (regression, binary classification, etc.)
- A target column (ground truth labels)
- An output column (model predictions)
- Model feature columns
- A baseline dataset

Update the artifact

DATASET_NAME = 'YOUR_DATASET_NAME'

dataset = fdl.Dataset.from_name(name=DATASET_NAME, model_id=model.id)

job = model.add_surrogate(
    dataset_id=dataset.id
)
job.wait()

Overview

Platform

Agentic AI Monitoring

LLM Monitoring

ML Monitoring

Experiments

Guardrails

Cookbooks

Tutorials

Client Library Reference

Surrogate Models

Surrogate Models

Surrogate Model prerequisites:

Update the artifact

​Surrogate Models

​Surrogate Model prerequisites:

​Update the artifact

Surrogate Models

Surrogate Model prerequisites:

Update the artifact