Sign in to view source links and access this dataset
Description
Iranian medical board examinations from 14 years provide 20,785 multiple-choice questions covering 23 specialties. The dataset, created by MohammadJRanjbar and updated in June 2025, includes 14,549 training, 1,000 validation, and 5,236 test items. Approximately 70% are clinical case scenarios, with the remainder covering basic science, and items are provided in both original Persian and an automatically translated English version.
Use Cases
Benchmarking LLM performance on Persian medical knowledge based on the 20,785 expert-validated questions.
Training specialized medical QA models using the 14,549 training items covering 23 specialties.
Evaluating model reasoning on clinical case scenarios, which constitute roughly 70% of the dataset.
Studying cross-lingual medical NLP by leveraging the parallel Persian and English question texts.
Analyzing question difficulty and patterns across medical specialties based on the source board exams.
Strengths
Large scale with 20,785 total expert-validated items.
Structured split into 14,549 training, 1,000 validation, and 5,236 test items.
Covers 23 distinct medical specialties for broad domain coverage.
Includes both original Persian text and an automatically translated English version.
Limitations
Column-level documentation is absent; field semantics must be inferred after download.
Freshness should be verified as the last metadata update was 2025-06-08.
Data may reflect geographic and institutional bias inherent to the source Iranian board exams.
Provenance
Source
MohammadJRanjbar on Hugging Face.
Collection Method
Collected from 14 years of Iranian residency and pre-residency board examinations.
Time Range
14 years of examination data (specific years not provided).
Freshness
Last updated 2025-06-08 17:11:34.
Geography
Iran.
License is unknown; terms of use must be verified before application.