TOFU
Benchmarks
| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Machine Unlearning | TOFU (5%) | Forget Quality86.6 | 45 | |
| Machine Unlearning | TOFU | Forget Quality (FQ)1 | 43 | |
| Knowledge Recovery | TOFU 10% (400 samples) 1.0 (forget) | ASR60 | 42 | |
| LLM Unlearning | TOFU 5% (forget set) | FQ-13.591 | 25 | |
| Machine Unlearning | TOFU 1.0 (Retain Set) | ROUGE-L100 | 24 | |
| Machine Unlearning | TOFU (10%) | Forget Quality (FQ)1 | 23 | |
| Machine Unlearning | TOFU 5% forget | Loss0.1434 | 20 | |
| Machine Unlearning | TOFU Llama-3.2 Instruct (average of 1%, 5%, 10% forget sets) | FQ Score-81.257 | 18 | |
| Machine Unlearning | TOFU 10% forget set 1.0 | FQ-238.973 | 18 | |
| Machine Unlearning | TOFU 1% (forget set) | FQ-2.17 | 18 | |
| Machine Unlearning | TOFU (1%) | Forget Quality (FQ)0.001 | 15 | |
| Machine Unlearning Evaluation | TOFU (Forget) | ROUGE0.986 | 14 | |
| Machine Unlearning Evaluation | TOFU Real World | ROUGE89.7 | 14 | |
| Machine Unlearning Evaluation | TOFU (Real Authors) | ROUGE93.3 | 14 | |
| Machine Unlearning | TOFU Llama 3.1 8B (5%) | FQ0.914 | 12 | |
| Machine Unlearning Evaluation | TOFU | Model Utility61.4 | 12 | |
| Information Extraction | TOFU 1% forget | Cumulative ROUGE-154 | 12 | |
| Machine Unlearning | TOFU Llama 3.1 8B (10%) | Forgetting Quality0.671 | 11 | |
| Machine Unlearning | TOFU | Forget Quality0 | 10 | |
| Machine Unlearning | TOFU Llama-3-3B (unlearning) | Extraction Strength6.042 | 10 | |
| Unlearning | ToFU | Probability0.9908 | 10 | |
| Unlearning | TOFU (forget01) | Forgetting Quality (F.Q.)92 | 10 | |
| Machine Unlearning | TOFU World Fact 1.0 | ROUGE-L0.88 | 10 | |
| Machine Unlearning | TOFU 1.0 (Real Author) | ROUGE-L93 | 10 | |
| Machine Unlearning | TOFU 1.0 (Forget Set) | ROUGE-L1 | 10 |
Showing 25 of 32 rows