jdh-algo/

MedDocBench

A compact benchmark of text‑rich medical document understanding covering routine, patient-uploaded artifacts from online consultations, spanning laboratory test reports (LTR) and general medical documents (GMD).

Configuration: v20251015-bytes Available splits: LTR_fullparsing: 100 LTR_simpleQA: 200 LTR_abnormalityQA (complex QA): 100 GMD_simpleQA: 100 GMD_complexQA: 100

Total: 600 QA pairs