Text-to-Image Generation on bias-prone prompts Average
0.582FairnessFairFlow
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| FairFlowBase Model=Stable Diffusion 32026.07 | 0.582 | 0.693 | 0.715 | 0.727 | |
| FairFlowBase Model=FLUX.1-dev2026.07 | 0.548 | 0.698 | 0.721 | 0.703 | |
| FairDiffusionBase Model=Stable Diffusion 32026.07 | 0.523 | 0.574 | 0.488 | 0.608 | |
| FairDiffusionBase Model=FLUX.1-dev2026.07 | 0.344 | 0.68 | 0.741 | 0.509 | |
| WeakGuidanceBase Model=FLUX.1-dev2026.07 | 0.342 | 0.666 | 0.73 | 0.505 | |
| WeakGuidanceBase Model=Stable Diffusion 32026.07 | 0.327 | 0.695 | 0.745 | 0.492 | |
| BiasNeuronBase Model=FLUX.1-dev2026.07 | 0.248 | 0.682 | 0.73 | 0.395 | |
| FairImagenBase Model=FLUX.1-dev2026.07 | 0.223 | 0.705 | 0.728 | 0.364 | |
| FairImagenBase Model=Stable Diffusion 32026.07 | 0.188 | 0.684 | 0.718 | 0.315 | |
| VanillaBase Model=Stable Diffusion 32026.07 | 0.148 | 0.707 | 0.749 | 0.258 | |
| BiasNeuronBase Model=Stable Diffusion 32026.07 | 0.137 | 0.71 | 0.738 | 0.241 | |
| VanillaBase Model=FLUX.1-dev2026.07 | 0.128 | 0.712 | 0.735 | 0.226 |