Skip to main content

Conversational G-Eval

The conversationl G-Eval is an adopted version of deepeval's popular GEval metric but for evaluating entire conversations instead. It is currently the best way to define custom critera to evaluate multi-turn conversations in deepeval. By defining a custom ConversationalGEval, you can easily determine whether your LLM chatbot is able to consistently generate responses that are up to standard with your custom criteria throughout a conversation.

Required Arguments​

To use the ConversationalGEval metric, you'll have to provide the following arguments when creating an ConversationalTestCase:

  • turns

Additionally, each LLMTestCases in turns requires the following arguments:

  • input
  • actual_output

You'll also need to supply any additional arguments such as expected_output and context if your evaluation criteria depends on these parameters.

Example​

To create a custom metric that evaluates entire LLM conversations, simply instantiate a ConversationalGEval class and define an evaluation criteria in everyday language:

from deepeval.test_case import LLMTestCase, LLMTestCaseParams, ConversationalTestCase
from deepeval.metrics import ConversationalGEval

convo_test_case = ConversationalTestCase(
    turns=[LLMTestCase(input="...", actual_output="...")]
)
professionalism_metric = ConversationalGEval(
    name="Professionalism",
    criteria="""Given the 'actual output' are generated responses from an
    LLM chatbot and 'input' are user queries to the chatbot, determine whether
    the chatbot has acted professionally throughout a conversation.""",
    # NOTE: you can only provide either criteria or evaluation_steps, and not both
    evaluation_steps=[
        "Check whether each LLM 'actual output' is professional with regards to the user 'input'",
        "Being professional means no profanity, no toxic language, and consistently says 'please' or 'thank you'.",
        "Penalize heavily if exclaimation marks are used in a rude demeanour."
    ],
    evaluation_params=[LLMTestCaseParams.INPUT, LLMTestCaseParams.ACTUAL_OUTPUT],
)

metric.measure(convo_test_case)
print(metric.score)
print(metric.reason)

There are three mandatory and six optional parameters required when instantiating an ConversationalGEval class:

  • name: name of metric. This will not affect the evaluation.
  • criteria: a description outlining the specific evaluation aspects for each test case.
  • evaluation_params: a list of type LLMTestCaseParams. Include only the parameters that are relevant for evaluation.
  • [Optional] evaluation_steps: a list of strings outlining the exact steps the LLM should take for evaluation. If evaluation_steps is not provided, ConversationalGEval will generate a series of evaluation_steps on your behalf based on the provided criteria. You can only provide either evaluation_steps OR criteria, and not both.
  • [Optional] threshold: the passing threshold, defaulted to 0.5.
  • [Optional] model: a string specifying which of OpenAI's GPT models to use, OR any custom LLM model of type DeepEvalBaseLLM. Defaulted to 'gpt-4o'.
  • [Optional] strict_mode: a boolean which when set to True, enforces a binary metric score: 1 for perfection, 0 otherwise. It also overrides the current threshold and sets it to 1. Defaulted to False.
  • [Optional] async_mode: a boolean which when set to True, enables concurrent execution within the measure() method. Defaulted to True.
  • [Optional] verbose_mode: a boolean which when set to True, prints the intermediate steps used to calculate said metric to the console, as outlined in the How Is It Calculated section. Defaulted to False.
danger

For accurate and valid results, only test case parameters that are mentioned in criteria/evaluation_steps should be included as a member of evaluation_params.

How Is It Calculated?​

The ConversationalGEval is an adapted version of GEval, so alike GEval, the ConversationalGEval metric is a two-step algorithm that first generates a series of evaluation_steps using chain of thoughts (CoTs) based on the given criteria, before using the generated evaluation_steps to determine the final score using the evaluation_params presented in an LLMTestCase of each turn.

Unlike regular GEval though, the ConversationalGEval takes the entire conversation history into account during evaluation.

tip

Similar to the original G-Eval paper, the ConversationalGEval metric uses the probabilities of the LLM output tokens to normalize the score by calculating a weighted summation. This step was introduced in the paper to minimize bias in LLM scoring, and is automatically handled by deepeval (unless you're using a custom LLM).