Guardrail Erosion in Domain Fine-Tuned Large Language Models: A Paired Audit of Knowledge-Centric Abstention in Biomedical LLMs
Abstract page for arXiv paper 2611.08780: Guardrail Erosion in Domain Fine-Tuned Large Language Models: A Paired Audit of Knowledge-Centric Abstention in Biomedical LLMs
arXiv.org