[2606.03648] Safety Measurements for Fine-tuned LLMs Should be Grounded in Capability
![[2606.03648] Safety Measurements for Fine-tuned LLMs Should be Grounded in Capability](/static/img/curated/desks/afbytes-us.png)
AI disclosure
Summary
Abstract page for arXiv paper 2606.03648: Safety Measurements for Fine-tuned LLMs Should be Grounded in Capability