Evaluating Retrieval-Augmented Generation vs. Long-Context Input for Clinical Reasoning over EHRs

Publication
arXiv preprint arXiv:2508.14817