Language Modeling on BERT Language Modeling c-pos non-masked (test)
0.9412AccuracyVanilla
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| VanillaIntervention=None2025.06 | 0.9412 | — | |
| Rand. Mean ProjectionIntervention=Random, Technique=Mean Projection2025.06 | 0.9405 | — | |
| Rand. LEACEIntervention=Random, Technique=LEACE2025.06 | 0.9395 | — | |
| Rand. INLP (repr)Intervention=Random, Technique=INLP, Source=Reproduced2025.06 | 0.9021 | — | |
| Amnesic LEACEIntervention=Amnesic, Technique=LEACE2025.06 | 0.9018 | 0.19 | |
| Rand. INLP (Elazar)Intervention=Random, Technique=INLP, Source=Elazar et al. (2021)2025.06 | 0.8965 | 0.36 | |
| Amnesic Mean ProjectionIntervention=Amnesic, Technique=Mean Projection2025.06 | 0.8687 | 0.49 | |
| Amnesic INLP (repr)Intervention=Amnesic, Technique=INLP, Source=Reproduced2025.06 | 0.7295 | 1.99 | |
| Amnesic INLP (Elazar)Intervention=Amnesic, Technique=INLP, Source=Elazar et al. (2021)2025.06 | 0.6192 | 3.21 |