Language Model Debiasing on StereoSet (test)
0.8535LMS ScoreRobustDebias
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| RobustDebiasBackbone=BERT, Multi-demographic Training=Yes (single model across demographics)2026.01 | 0.8535 | 0.5225 | 0.8151 | |
| FineDebBackbone=BERT, Multi-demographic Training=No (specific models per demographic)2026.01 | 0.8523 | 0.5465 | 0.773 | |
| PCGUBackbone=BERT2026.01 | 0.8471 | 0.5375 | 0.7836 | |
| CausalDebiasBackbone=BERT, Multi-demographic Training=No (specific models per demographic)2026.01 | 0.7485 | 0.5291 | 0.7049 | |
| AutoDebiasBackbone=BERT, Multi-demographic Training=No (specific models per demographic)2026.01 | 0.6371 | 0.5316 | 0.5968 |