langtest.transform.fairness.MinGenderLLMEval#

class MinGenderLLMEval#

Bases: BaseFairness

Class for evaluating fairness based on minimum gender performance in question-answering tasks using a Language Model.

alias_name#

Alias names for the evaluation method.

Type:

List[str]

supported_tasks#

Supported tasks for this evaluation method.

Type:

List[str]

transform(cls, test

str, data: List[Sample], params: Dict) -> List[MaxScoreSample]: Transforms data for evaluation.

run(sample_list

List[MaxScoreSample], grouped_label: Dict[str, Tuple[List, List]], **kwargs) -> List[MaxScoreSample]: Runs the evaluation process.

__init__()#

Methods

__init__()

async_run(sample_list, model, **kwargs)

Creates a task for the run method.

run(sample_list, grouped_label, **kwargs)

Runs the evaluation process using Language Model.

transform(test, data, params)

Transforms the data for evaluation.

Attributes

alias_name

eval_model

supported_tasks

test_types

async classmethod async_run(sample_list: List[Sample], model: ModelAPI, **kwargs)#

Creates a task for the run method.

Parameters:

  • sample_list (List[Sample]) – The input data to be transformed.

  • model (ModelAPI) – The model to be used for the computation.

Returns:

The task for the run method.

Return type:

asyncio.Task

async static run(sample_list: List[MaxScoreSample], grouped_label, **kwargs) → List[MaxScoreSample]#

Runs the evaluation process using Language Model.

Parameters:

  • sample_list – The input data samples.

  • grouped_label – A dictionary containing grouped labels where each key corresponds to a test case and the value is a tuple containing true labels and predicted labels.

  • **kwargs – Additional keyword arguments.

Returns:

The evaluated data samples.

classmethod transform(test: str, data: List[Sample], params: Dict) → List[MaxScoreSample]#

Transforms the data for evaluation.

Parameters:

  • test (str) – The test alias name.

  • data (List[Sample]) – The data to be transformed.

  • params (Dict) – Parameters for transformation.

Returns:

The transformed data samples.

Return type:

List[MaxScoreSample]