Skip to main content
. 2024 Sep 9;111(10):2190–2202. doi: 10.1016/j.ajhg.2024.08.010

Table 1.

Datasets used for LLM evaluation in this study

Source # of individuals with HPO terms as input # of individuals with narratives as input Average # of HPO (s.d.) Average # of tokens (s.d.)
AJHG 78 72 11.42 (6.43) 187.64 (88.87)
CSH 72 49 12.83 (8.30) 161.43 (102.38)
ColumbiaU 27 11.52 (4.73)
DGD 85 9.03 (4.06)
TAF1 14 34.86 (10.71)
Other 4 180.25 (143.98)
Post2023 130 270.09 (177.07)