FairnessConfig
Options & Configuration · AiDotNet.Safety
Configuration for fairness and bias detection.
For Beginners
Fairness settings check whether the model treats all demographic groups equitably. For example, ensuring the model doesn’t give different quality results for different genders or ethnicities.
How It Works
References:
- BEATS: Comprehensive bias evaluation test suite for LLMs (2025)
- SB-Bench: Stereotype bias benchmark for multimodal models (2025)
- Demographic-targeted bias: race/ethnicity 55.6% exploitability (2025)
Constructors
| Constructor | Summary |
|---|---|
FairnessConfig() |
Properties
| Property | Summary |
|---|---|
bool? DemographicParity { get; set; } |
Gets or sets whether demographic parity checking is enabled. |
bool? EqualizedOdds { get; set; } |
Gets or sets whether equalized odds checking is enabled. |
bool? StereotypeDetection { get; set; } |
Gets or sets whether stereotype detection is enabled. |
string[] ProtectedAttributes { get; set; } |
Gets or sets the protected attributes to monitor. |