deep-ignorance-mu-sft-ret140-up1
Unlearned model using max update baseline (SFT, full-rank).
Base model: EleutherAI/deep-ignorance-unfiltered
Training hyperparameters
| Parameter | Value |
|---|---|
| Algorithm | Max Update |
| Mode | SFT (full-rank) |
| retain_coef | 140 |
| update_coef | 1 |
| Learning rate | 5e-5 |
Evaluation results
| Benchmark | Accuracy |
|---|---|
| WMDP Bio Robust | 0.3594 |
| MMLU | 0.4143 |
- Downloads last month
- 17
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support
Model tree for EleutherAI/deep-ignorance-mu-sft-ret140-up1
Unable to build the model tree, the base model loops to the model itself. Learn more.