Open-Source AI Auditing Libraries
The libraries behind GPTfake’s published data, released so anyone can reproduce our findings or run their own audits. All three are MIT-licensed and developed in the open on github.com/gptfake .
Bias Detector
Score political and cultural bias in AI responses.
Censorship TrackerLog refusal patterns over time and detect policy drift.
Transparency AnalyzerGrade how transparently a model explains its refusals.
Each library applies the same monitoring methodology GPTfake uses to produce its live model data and quarterly reports. Looking for developer tooling instead? See the CLI and browser extension.
GPTfake