UnBias-Plus is an open-source toolkit designed to address persistent bias in natural language by unifying detection, explanation, and neutral rewriting capabilities.
- Performs segment-level multi-class bias classification and biased span localization.
- Generates reasoning for each decision to provide interpretable explanations.
- Offers neutral text rewriting to mitigate identified biases.
- Accessible via Python, CLI, REST API, and web interfaces with publicly available models and datasets.