Measuring the sensitivity of LLM-based structured extraction to prompt, model, and schema choices in clinical discharge summaries 事件

Measuring the sensitivity of LLM-based structured extraction to prompt, model, and schema choices in clinical discharge summaries arXiv:2606.05970v1 Announce Type: new Abstract: Large language models are increasingly used for structured extraction from clinical free-text notes, but the sensitivity of their output to upstream configuration choices is less understood than their accuracy on fixed benchmarks. This work measures that sensitivity without human-annotated ground truth, by holding the e