BiasScope is a Python library for measuring bias in language models across four complementary families of metrics:

• embedding-based metrics
• probability-based metrics
• generated-text metrics
• prompt-based benchmarks

The goal is a single, consistent API for bias evaluation whether you are working with sentence encoders, masked language models, generated completions, or dataset-driven benchmark suites.

• One package for multiple bias evaluation paradigms
• Consistent metric classes with evaluate() entrypoints
• Optional model adapters so users do not need to hand-wire every scorer
• Support for both raw-text convenience paths and precomputed inputs where appropriate
• Lightweight core install with optional extras for heavier ML dependencies

Core install:

pip install bias-scope

Optional extras:

pip install "bias-scope[torch]"
pip install "bias-scope[embeddings]"
pip install "bias-scope[datasets]"
pip install "bias-scope[llm]"
pip install "bias-scope[all]"

What each extra includes:

• torch: torch, transformers for probability-based masked-token metrics, BertPLLScorer, and transformer-backed generated-text metrics such as RegardScore
• embeddings: sentence-transformers for the built-in embedding helper used by embedding-based convenience paths
• datasets: datasets for prompt-based benchmark loaders
• llm: litellm for prompt-based model calls
• all: everything above

Install from source:

git clone https://github.com/RAINLabLAU/bias_scope.git
cd bias_scope
pip install -e .

You can now pass raw text directly and let the metric embed it for you:

from bias_scope.embeddings_based import WEAT

weat = WEAT(model_name="sentence-transformers/all-MiniLM-L6-v2")
score = weat.evaluate(
    target_embeddings=(
        ["John", "Paul", "Mike", "Kevin"],
        ["Amy", "Joan", "Lisa", "Sarah"],
    ),
    attribute_embeddings=(
        ["executive", "management", "salary", "career"],
        ["home", "children", "marriage", "family"],
    ),
)
print(f"WEAT effect size: {score:.4f}")

If you already have embeddings, you can still pass precomputed arrays exactly as before.

Masked-token metrics can use a built-in model adapter via model_name:

from bias_scope.probability_based import CrowSPairs

crows = CrowSPairs(model_name="bert-base-uncased")
score = crows.evaluate(
    sentence_pairs=[
        (
            ["Women", "are", "bad", "at", "math"],
            ["Men", "are", "bad", "at", "math"],
        )
    ]
)
print(f"CrowS-Pairs score: {score:.4f}")

Advanced users can still provide a custom callback or scorer wrapper when needed.

from bias_scope.generated_text_based import ScoreParity

parity = ScoreParity(classifier=lambda texts: [0.9 if "doctor" in t else 0.4 for t in texts])
result = parity.evaluate(
    group_a_texts=[["The man is a doctor."]],
    group_b_texts=[["The woman is a nurse."]],
)
print(result)

from bias_scope.prompts_based import BBQMetric

metric = BBQMetric(model_name="gpt-4o-mini")
result = metric.evaluate(return_details=True)
print(result)

Prompt-based metrics typically require bias-scope[datasets], bias-scope[llm], or both depending on the benchmark.

Use these when you want to measure association bias in vector spaces.

• WEAT
• SEAT
• CEAT
• SentenceBiasScore
• embed() helper for built-in text embedding

Use these with masked or token-prediction models.

• CrowSPairs
• AUL
• AULA
• CAT
• ICAT
• LMB
• LPBS
• CBS
• DisCoMetric
• BertPLLScorer
• TokenPredictionScorer

Use these when you already have generations or want to score generated completions.

• ToxicityFraction
• ToxicityProbability
• RegardScore
• ScoreParity
• SocialGroupSubstitution
• CoOccurrenceBiasScore
• CounterfactualSentimentBias
• DemographicRepresentation
• StereotypicalAssociations
• MarkedPersons
• EMT
• FGB
• GenderPolarity
• HONEST
• PGB
• PerspectiveAPIClient
• PsycholinguisticNorms

Use these for dataset-backed evaluation suites and benchmark-style audits.

• AnalogicalReasoningBias
• BBQMetric
• BOLD
• CounterfactualFairness
• DemographicRepresentationBias
• OpinionConsistencyAcrossPersonas
• RealToxicityPrompts
• StereoSetMetric
• TofNof
• TruthfulQA
• UnQoverMetric

• Most metrics return a scalar by default.
• Metrics that support return_details=True return a richer dictionary of component scores.
• Embedding metrics accept model_name on both __init__ and evaluate(); the per-call value overrides the instance default.
• Probability-based masked-token metrics support either a built-in model_name path or a backward-compatible custom callback path.
• Metric objects now have informative repr(...) output for notebook and REPL use.

The repository includes runnable examples for each metric family:

• examples/embeddings_based
• examples/probability_based
• examples/generated_text_based
• examples/prompts_based
• examples/metric_usage_examples.py

Project docs live under docs/.

Good starting points:

• docs/getting-started/installation.md
• docs/getting-started/quickstart.md
• docs/api

Install developer dependencies:

pip install -e .[dev]

Run tests:

python -m pytest

This project is licensed under the MIT License.