OpenAudit
VerifiedOpen-source framework for detecting algorithmic bias in LLMs using established audit study methodologies. Supports 26+ models including GPT, Claude, and Gemini.
Frameworks Covered
LL144EU-AI-Act
Services
bias-auditllm-bias-testingstatistical-analysishiring-bias
About OpenAudit
OpenAudit is a production-ready framework for detecting algorithmic bias in Large Language Models. It provides empirical bias testing with real LLM responses and statistical analysis for AI fairness research, using established audit study methodologies.
Features
- Multi-LLM integration (26+ models: GPT, Claude, Gemini, Grok, Llama, Deepseek)
- Real-time bias detection from actual LLM responses
- CV generation for controlled hiring bias experiments
- Statistical analysis with significance testing and effect sizes
- Ceteris paribus (all else equal) testing methodology
- Publication-quality bias analysis and reporting
Use Cases
- Hiring bias studies in AI systems
- Cross-model bias comparison research
- NYC LL144 AEDT bias audit evidence
- AI fairness policy research
Loading reviews...