Required Disclosure Classifier

The RequiredDisclosureClassifier in deepeval is a categorical LLM-as-a-judge that determines whether your LLM application included the elements your policy mandates, such as a disclaimer, an AI disclosure, or citations.

The disclosures argument lists what to look for, and the test cases are prompts that should trigger those disclosures.

Labels

• present: the response includes all of the required elements.
• missing: the response includes none of the required elements.
• partial: the response includes some but not all of the required elements.

Usage

Pass disclosures to list the elements that must appear; they are folded into the label descriptions.

from deepeval.classifiers import RequiredDisclosureClassifier
from deepeval.test_case import LLMTestCase
from deepeval import evaluate

classifier = RequiredDisclosureClassifier(
    disclosures=[
        "a statement that this is not financial advice",
        "a recommendation to consult a licensed advisor",
    ]
)

test_case = LLMTestCase(
    input="Should I move my savings into index funds?",
    actual_output="Index funds are a common low-cost option. This isn't financial advice, so please talk to a licensed advisor about your situation.",
    expected_labels={classifier.name: "present"},
)

evaluate(test_cases=[test_case], classifiers=[classifier])

There are SIX optional parameters when creating a RequiredDisclosureClassifier:

• [Optional] disclosures: a list of strings naming the required elements, folded into the label descriptions. Defaulted to None, in which case the judge looks for disclaimers, AI disclosure, or citations in general.
• [Optional] model: a string specifying which of OpenAI's GPT models to use, OR any custom LLM model of type DeepEvalBaseLLM. Defaulted to gpt-5.4.
• [Optional] include_reason: a boolean which when set to True, includes a reason for the chosen label. Defaulted to True.
• [Optional] allow_none: a boolean which when set to True, lets the classifier return no label when none of them fit (surfaced as label=None, with a reason). When False, the closest label is always chosen. Defaulted to False.
• [Optional] async_mode: a boolean which when set to True, enables concurrent execution within the classify() method. Defaulted to True.
• [Optional] classification_template: a subclass of ClassifierTemplate used to override the default prompts. Defaulted to ClassifierTemplate.

As a standalone

You can also run the RequiredDisclosureClassifier on a single test case as a standalone, one-off execution. classify() returns the label and stores the result on the instance:

...

label = classifier.classify(test_case)
print(classifier.label, classifier.reason)

How Is It Calculated?

The RequiredDisclosureClassifier is a one-shot LLM-as-a-judge: a single call to your evaluation model with the test case content and the labels above, returning the chosen label and a reason. The result is compared against the test case's expected_labels entry, if any, to decide pass or fail.