ys-zong/VLGuard
[ICML 2024] Safety Fine-Tuning at (Almost) No Cost: A Baseline for Vision Large Language Models.
VLGuard helps AI researchers and developers working with Vision Large Language Models (VLMs) to make them safer and more helpful. It provides a dataset and pre-trained model weights that can be used to fine-tune existing VLMs. This results in models that are less likely to generate harmful content while maintaining their ability to assist users.
No commits in the last 6 months.
Use this if you are building or deploying Vision Large Language Models and need to improve their safety performance efficiently without sacrificing helpfulness.
Not ideal if you are looking for a ready-to-use, end-user application rather than tools for VLM development and fine-tuning.
Stars
85
Forks
5
Language
Python
License
—
Category
Last pushed
Jan 19, 2025
Commits (30d)
0
Get this data via API
curl "https://pt-edge.onrender.com/api/v1/quality/transformers/ys-zong/VLGuard"
Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.
Higher-rated alternatives
steering-vectors/steering-vectors
Steering vectors for transformer language models in Pytorch / Huggingface
jianghoucheng/AlphaEdit
AlphaEdit: Null-Space Constrained Knowledge Editing for Language Models, ICLR 2025 (Outstanding Paper)
kmeng01/memit
Mass-editing thousands of facts into a transformer memory (ICLR 2023)
boyiwei/alignment-attribution-code
[ICML 2024] Assessing the Brittleness of Safety Alignment via Pruning and Low-Rank Modifications
jianghoucheng/AnyEdit
AnyEdit: Edit Any Knowledge Encoded in Language Models, ICML 2025