StudyTypeTeller-Large language models to automatically classify research study types for systematic reviews.

Despite overlapping confidence intervals, there was a clear trend that SciBERT also outperformed GPT at class-level for most classes, except that GPT-4 outperformed SciBERT for Human-RCT-drug-intervention , Animal-non-drug-intervention, and Animal-drug-intervention (Figure 2 and Figure S2 in the Supplementary Material). 3.5 Comparison to related work The MeSH-based approach performed well in capturing nearly all animal articles (high recall).

also in 13,370 other papers