Instructions to use xudongwu/SafeSelfPlay-checkpoints with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- PEFT
How to use xudongwu/SafeSelfPlay-checkpoints with PEFT:
Task type is invalid.
- Notebooks
- Google Colab
- Kaggle
SafeSelfPlay checkpoints
lora/A1 through lora/D3 are the canonical role-specific PEFT adapters.
self-redteam-reproduction/step200 is our reproduction of public Self-RedTeam
commit 0c56e503e8ae1b1b0fcd2214c92ea31fef1cb123; it is not an
author-released checkpoint. The authors' weights are available in the
official collection.
Training and loading commands are documented in SafeSelfPlay.
- Downloads last month
- -
Model tree for xudongwu/SafeSelfPlay-checkpoints
Base model
meta-llama/Llama-3.1-8B Finetuned
meta-llama/Llama-3.1-8B-Instruct