Small LLMs for Biomedical Claim Verification: Cost-Effective Fine-Tuning, Structural Dataset Shortcuts, and Cross-Domain Generalization
用小语言模型低成本搞定生物医学声明验证,揭秘结构化数据集捷径与跨域泛化新发现。
arXiv:2606.12854v1 Announce Type: new Abstract: Large Language Models such as GPT-4o and GPT-5 achieve strong zero-shot performance on biomedical clai…