4.2. Assessing performance: Metrics that Matter

With an evaluation set, you are able to measure the performance of your agent application across a number of different dimensions, including:

• Retrieval quality: Retrieval metrics assess how successfully your agent application retrieves relevant supporting data. Precision and recall are two key retrieval metrics.

• Response quality: Response quality metrics assess how well the agent application responds to a user’s request. Response metrics can measure, for instance, if the resulting answer is accurate per the ground-truth, how well-grounded the response was given the retrieved context (e.g., did the LLM hallucinate), or how safe the response was (e.g., no toxicity).

• System performance (cost & latency): Metrics capture the overall cost and performance of agent applications. Overall latency and token consumption are examples of agent performance metrics.

It is very important to collect both response and retrieval metrics. an agent application can respond poorly in spite of retrieving the correct context; it can also provide good responses on the basis of faulty retrievals. Only by measuring both components can we accurately diagnose and address issues in the application.

There are two key approaches to measuring performance across these metrics:

• Deterministic measurement: Cost and latency metrics can be computed deterministically based on the application’s outputs. If your evaluation set includes a list of documents that contain the answer to a question, a subset of the retrieval metrics can also be computed deterministically.

• LLM judge based measurement In this approach, a separate LLM acts as a judge to evaluate the quality of the agent application’s retrieval and responses. Some LLM judges, such as answer correctness, compare the human-labeled ground truth vs. the app’s outputs. Other LLM judges, such as groundedness, do not require human-labeled ground truth to assess their app’s outputs.

Important

For an LLM judge to be effective, it must be tuned to understand the use case. Doing so requires careful attention to understand where the judge does and does not work well, and then tuning the judge to improve it for the failure cases.

Mosaic AI Agent Evaluation provides an out-of-the-box implementation, using hosted LLM judge models, for each metric discussed on this page. Agent Evaluation’s documentation discusses the details of how these metrics and judges are implemented and provides capabilities to tune the judge’s with your data to increase their accuracy

4.2.1. Metric overview

Below is a summary of the metrics that Databricks recommends for measuring the quality, cost, and latency of your agent application. These metrics are implemented in Mosaic AI Agent Evaluation.

Dimension	Metric name	Question	Measured by	Ground truth required?
Retrieval	chunk_relevance/precision	What % of the retrieved chunks are relevant to the request?	LLM judge	✖️
Retrieval	document_recall	What % of the ground truth documents are represented in the retrieved chunks?	Deterministic	✔️
Response	correctness	Overall, did the agent generate a correct response?	LLM judge	✔️
Response	relevance_to_query	Is the response relevant to the request?	LLM judge	✖️
Response	groundedness	Is the response a hallucination or grounded in context?	LLM judge	✖️
Response	safety	Is there harmful content in the response?	LLM judge	✖️
Cost	total_token_count, total_input_token_count, total_output_token_count	What's the total count of tokens for LLM generations?	Deterministic	✖️
Latency	latency_seconds	What's the latency of executing the app?	Deterministic	✖️

4.2.1.1. Understanding how retrieval metrics work

Retrieval metrics help you understand if your retriever is delivering relevant results. Retrieval metrics are based on precision and recall.

Metric Name	Question Answered	Details
Precision	What % of the retrieved chunks are relevant to the request?	Precision is the proportion of retrieved documents that are actually relevant to the user’s request. An LLM judge can be used to assess the relevance of each retrieved chunk to the user’s request.
Recall	What % of the ground truth documents are represented in the retrieved chunks?	Recall is the proportion of the ground truth documents that are represented in the retrieved chunks. This is a measure of the completeness of the results.

Below is a quick primer on Precision & Recall adapted from the excellent Wikipedia article.

Precision measures “Of the chunks I retrieved, what % of these items are actually relevant to my user’s query?” Computing precision does NOT require knowing ALL relevant items.

Recall measures “Of ALL the document that I know are relevant to my user’s query, what % did I retrieve a chunk from?” Computing recall requires your ground-truth to contain ALL relevant items.

* Items can either be a document or a chunk of a document.

In the example below, two out of the three retrieved results were relevant to the user’s query, so the precision was 0.66 (2/3). The retrieved docs included two out of a total of four relevant docs, so the recall was 0.5 (2/4).