Answer and Context Relevance Judges

MLflow provides two built-in LLM judges to assess relevance in your GenAI applications. These judges help diagnose quality issues - if context isn't relevant, the generation step cannot produce a helpful response.

• RelevanceToQuery: Evaluates if your app's response directly addresses the user's input
• RetrievalRelevance: Evaluates if each document returned by your app's retriever(s) is relevant

Prerequisites for running the examples

1. Install MLflow and required packages

   bash

   pip install --upgrade mlflow

2. Create an MLflow experiment by following the setup your environment quickstart.

3. (Optional, if using OpenAI models) Use the native OpenAI SDK to connect to OpenAI-hosted models. Select a model from the available OpenAI models.

   python

   import mlflow
   import os
   import openai
   
   # Ensure your OPENAI_API_KEY is set in your environment
   # os.environ["OPENAI_API_KEY"] = "<YOUR_API_KEY>" # Uncomment and set if not globally configured
   
   # Enable auto-tracing for OpenAI
   mlflow.openai.autolog()
   
   # Create an OpenAI client
   client = openai.OpenAI()
   
   # Select an LLM
   model_name = "gpt-4o-mini"

Usage Examples

RelevanceToQuery Judge

This judge evaluates if your app's response directly addresses the user's input without deviating into unrelated topics.

You can invoke the judge directly with a single input for testing, or pass it to mlflow.genai.evaluate for running full evaluation on a dataset.

Requirements:

• Trace requirements: inputs and outputs must be on the Trace's root span

Invoke directly

python

import mlflow
from mlflow.genai.scorers import RelevanceToQuery

assessment = RelevanceToQuery(name="my_relevance_to_query")(
    inputs={"question": "What is the capital of France?"},
    outputs="The capital of France is Paris.",
)
print(assessment)

Invoke with evaluate()

python

import mlflow
from mlflow.genai.scorers import RelevanceToQuery

data = [
    {
        "inputs": {"question": "What is the capital of France?"},
        "outputs": "The capital of France is Paris.",
    }
]
result = mlflow.genai.evaluate(data=data, scorers=[RelevanceToQuery()])

RetrievalRelevance Judge

This judge evaluates if each document returned by your app's retriever(s) is relevant to the input request. It evaluates each retriever span separately and returns a separate Feedback object for each retriever span in your trace.

Requirements:

• Trace requirements: The MLflow Trace must contain at least one span with span_type set to RETRIEVER

Invoke directly

python

from mlflow.genai.scorers import RetrievalRelevance
import mlflow

# Get a trace from a previous run
trace = mlflow.get_trace("<your-trace-id>")

# Assess if each retrieved document is relevant
feedbacks = RetrievalRelevance()(trace=trace)
print(feedbacks)

Invoke with evaluate()

python

import mlflow
from mlflow.genai.scorers import RetrievalRelevance

# Evaluate traces from previous runs
results = mlflow.genai.evaluate(
    data=traces,  # DataFrame or list containing trace data
    scorers=[RetrievalRelevance()],
)

tip

For a complete RAG application example with these judges, see the RAG Evaluation guide.

Select the LLM that powers the judge

You can change the judge model by using the model argument in the judge definition. The model must be specified in the format <provider>:/<model-name>, where <provider> is a LiteLLM-compatible model provider.

For a list of supported models, see selecting judge models.

Interpret results

The judge returns a Feedback object containing:

• value: "yes" if context is relevant, "no" if not
• rationale: Explanation of why the judge found the context relevant or irrelevant

Next steps

Explore other built-in judges

Learn about groundedness, safety, and correctness judges

Learn more →

Create custom judges

Build specialized judges for your use case

Learn more →

Evaluate RAG applications

Apply relevance judges in comprehensive RAG evaluation

Learn more →