Sign in to view source links and access this dataset
Description
1,360 multiple-choice questions focusing on rare disease diagnosis. The dataset covers 205 types of rare diseases, with questions distributed across symptoms (11%), causes (33%), affected populations (13%), related disorders (15%), diagnosis (15%), and other properties (9%). It was created by guan-wang and last updated on the Hugging Face platform in August 2024.
Use Cases
Training or benchmarking medical question-answering models based on the multi-choice format.
Analyzing knowledge representation for rare diseases based on the described question categories (symptoms, causes, diagnosis).
Developing educational tools for medical professionals based on the diagnostic and clinical knowledge covered.
Strengths
Covers 205 distinct rare diseases, providing breadth in a specialized domain.
Contains 1,360 questions with a defined distribution across six clinical categories.
Provides a structured multi-choice format suitable for model evaluation.
Limitations
Column-level documentation is absent; field semantics must be inferred after download.
Row count is known, but other scale metrics like file size and format are unknown.
Data may reflect source bias inherent to its compilation method, which is not described.
Provenance
Source
Hugging Face dataset by author guan-wang.
Freshness
Last updated 2024-08-21 17:13:24; freshness should be verified.
License is unknown; users must verify terms of use before application.