tomg-group-umd

/

LoRI-D_safety_llama3_rank_64

like 0

Follow

Tom Goldstein's Lab at University of Maryland, College Park 181

This PR significantly improves the model card for LoRI-D_safety_llama3_rank_64 by:

• Adding license: apache-2.0 and tags: - lora to the metadata.
• Providing a comprehensive model description based on the paper abstract and GitHub README.
• Including explicit links to the GitHub repository and the Hugging Face collection of LoRI adapters.
• Adding a practical "How to Get Started" code snippet for direct use of the model, including a specific example for safety alignment.
• Filling in relevant sections like "Model Details", "Uses", "Bias, Risks, and Limitations", "Training Details", and "Evaluation", and removing empty placeholders.
• Including the architecture image from the GitHub repository.
• Populating the citation section with the BibTeX entry.

This makes the model card much more informative and user-friendly, adhering to Hugging Face Hub best practices.