tofu_Llama-3.1-8B-Instruct_forget01_SatImp
open-unlearning/tofu_Llama-3.1-8B-Instruct_full unlearned on the TOFU forget01 split with SatImp, trained with the open-unlearning framework. Used as a weight-unlearning baseline / draft model in the Speculative-Decoding-Unlearning project.
Full training config: .hydra/config.yaml. TOFU evaluation outputs: evals/.
Method hyperparameters
gamma: 1.0
alpha: 0.1
retain_loss_type: NLL
beta1: 5.0
beta2: 1.0
TOFU summary metrics
| metric | value |
|---|---|
| exact_memorization | 0.9406 |
| extraction_strength | 0.4152 |
| forget_Q_A_PARA_Prob | 0.0397 |
| forget_Q_A_gibberish | 0.8397 |
| forget_quality | 0.0286 |
| forget_truth_ratio | 0.5041 |
| mia_loss | 0.9719 |
| mia_min_k | 0.9706 |
| mia_min_k_plus_plus | 0.7956 |
| mia_zlib | 0.9788 |
| model_utility | 0.6110 |
| privleak | -94.1250 |