Text-to-Image Generation on bias-prone prompts (Gender)
0.648Fairness MetricFairDiffusion
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| FairDiffusionBase Model=Stable Diffusion 32026.07 | 0.648 | 0.583 | 0.484 | 0.687 | |
| FairFlowBase Model=FLUX.1-dev2026.07 | 0.6 | 0.704 | 0.725 | 0.746 | |
| FairFlowBase Model=Stable Diffusion 32026.07 | 0.597 | 0.696 | 0.739 | 0.744 | |
| WeakGuidanceBase Model=FLUX.1-dev2026.07 | 0.486 | 0.671 | 0.727 | 0.647 | |
| WeakGuidanceBase Model=Stable Diffusion 32026.07 | 0.3 | 0.697 | 0.748 | 0.461 | |
| BiasNeuronBase Model=FLUX.1-dev2026.07 | 0.276 | 0.67 | 0.726 | 0.429 | |
| FairDiffusionBase Model=FLUX.1-dev2026.07 | 0.244 | 0.685 | 0.74 | 0.391 | |
| FairImagenBase Model=FLUX.1-dev2026.07 | 0.236 | 0.706 | 0.728 | 0.381 | |
| FairImagenBase Model=Stable Diffusion 32026.07 | 0.173 | 0.682 | 0.719 | 0.293 | |
| BiasNeuronBase Model=Stable Diffusion 32026.07 | 0.128 | 0.71 | 0.738 | 0.227 | |
| VanillaBase Model=Stable Diffusion 32026.07 | 0.113 | 0.707 | 0.749 | 0.203 | |
| VanillaBase Model=FLUX.1-dev2026.07 | 0.084 | 0.712 | 0.735 | 0.155 |