Search papers, labs, and topics across Lattice.
University of Massachusetts
1
0
2
LLMs can misclassify harmless rephrasings as critical errors, but a new lightweight metric outperforms larger models in clinical significance evaluation.
