Search papers, labs, and topics across Lattice.
University of Delaware
2
0
3
16
F1 scores can be misleading, as 13 out of 15 LLMs fail to discriminate errors better than chance in clinical documentation despite appearing competent.
Clinical text summarization gets a boost in factual accuracy with AgenticSum, an agentic framework that selectively revises weakly supported spans identified via internal attention grounding.
