| title: Sequential Forgetting Leaderboard | |
| emoji: 📉 | |
| colorFrom: gray | |
| colorTo: red | |
| sdk: static | |
| app_file: index.html | |
| license: apache-2.0 | |
| short_description: Which fine-tuning methods forget least — with receipts | |
| # Sequential Forgetting Leaderboard | |
| Ranks continual fine-tuning approaches by how much they destroy previously | |
| learned tasks. Every number links to a raw run file; invalid runs are | |
| disclosed, not deleted. Data: the | |
| [sequential-forgetting-benchmark](https://huggingface.co/datasets/ModelBrew/sequential-forgetting-benchmark) | |
| dataset. Rebuild the page: `python3 build_static.py`. Maintained by [ModelBrew](https://modelbrew.ai). | |