LLM PMID Evaluation SemMedDB v1.0
LLM PMID Evaluation Results (SemMedDB)
This release contains the evaluation results of research triples from SemMedDB against PubMed abstracts using LLMs.
Files
The data is split into 5 parts due to GitHub's 2 GB file size limit. To reassemble:
cat LLM_Pmid_Evaluation_SemMedDB_with_names_v1.0.tar.gz.part_a{a,b,c,d,e} > LLM_Pmid_Evaluation_SemMedDB_with_names_v1.0.tar.gz
tar -xzvf LLM_Pmid_Evaluation_SemMedDB_with_names_v1.0.tar.gzContents
- results_with_names.parquet — Full evaluation results with abstracts, including subject/object names
- results_no_abstract_with_names.parquet — Evaluation results without abstract text (smaller file), including subject/object names