Enhancing privacy-preserving deployable large language models for perioperative complication detection: a targeted strategy with LoRA fine-tuning.

For the expanded Center 2 external validation cohort ( n = 102), all models demonstrated highly significant improvements with the comprehensive vs targeted approach ( p < 0.001).

also in 132,142 other papers

While most models demonstrated subtle improvement trends, the majority did not reach statistical significance for F1 and recall metrics (most comparisons n.s.).

also in 111,027 other papers