LangFair: Assess your LLM use case for bias and fairness risks

[dylanbouchard]

LangFair is an open-source Python package that equips LLM practitioners with tools to evaluate bias and fairness risks relevant to their specific use cases. The package offers functionality to easily generate evaluation datasets, comprised of LLM responses to use-case-specific prompts, and subsequently calculate applicable metrics for the practitioner's use case. To guide in metric selection, LangFair offers an actionable decision framework.

To generate evaluation datasets, LangFair is built on top of LangChain. For more information on LangFair, please see the below resources:

• Github repository
• Documentation site
• Software paper on arXiv