Lexical Gender Dataset evaluation on LGD
99.51CompletenessAlterRep
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| AlterRepModel=EleutherAI/pythia-70m2026.05 | 99.51 | 32.34 | 48.81 | |
| AlterRepModel=Meta-Llama-3-8B-Instruct2026.05 | 94.9 | 65.36 | 77.41 | |
| HDMIModel=Meta-Llama-3-8B-Instruct2026.05 | 94.12 | 81.17 | 87.16 | |
| HDMIModel=EleutherAI/pythia-70m2026.05 | 93.41 | 85.38 | 89.21 | |
| FGSMModel=Meta-Llama-3-8B-Instruct2026.05 | 58.13 | 33.37 | 42.4 | |
| PGDModel=Meta-Llama-3-8B-Instruct2026.05 | 54.02 | 41.49 | 46.94 | |
| PGDModel=EleutherAI/pythia-70m2026.05 | 54.02 | 41.49 | 46.94 | |
| FGSMModel=EleutherAI/pythia-70m2026.05 | 43.93 | 99.59 | 60.97 |