---
dataset_info:
- config_name: all
  features:
  - name: answer
    dtype: string
  - name: question
    dtype: string
  - name: difficulty
    dtype: string
  - name: question_type
    dtype: string
  - name: interrogative_type
    dtype: string
  - name: main_medical_field
    dtype: string
  - name: detailed_medical_field
    dtype: string
  - name: option_a
    dtype: string
  - name: option_b
    dtype: string
  - name: option_c
    dtype: string
  - name: option_d
    dtype: string
  splits:
  - name: test
    num_bytes: 1589543
    num_examples: 4918
  download_size: 622399
  dataset_size: 1589543
- config_name: easy
  features:
  - name: answer
    dtype: string
  - name: question
    dtype: string
  - name: difficulty
    dtype: string
  - name: question_type
    dtype: string
  - name: interrogative_type
    dtype: string
  - name: main_medical_field
    dtype: string
  - name: detailed_medical_field
    dtype: string
  - name: option_a
    dtype: string
  - name: option_b
    dtype: string
  - name: option_c
    dtype: string
  - name: option_d
    dtype: string
  splits:
  - name: test
    num_bytes: 711060.3090687271
    num_examples: 2200
  download_size: 233535
  dataset_size: 711060.3090687271
- config_name: hard
  features:
  - name: answer
    dtype: string
  - name: question
    dtype: string
  - name: difficulty
    dtype: string
  - name: question_type
    dtype: string
  - name: interrogative_type
    dtype: string
  - name: main_medical_field
    dtype: string
  - name: detailed_medical_field
    dtype: string
  - name: option_a
    dtype: string
  - name: option_b
    dtype: string
  - name: option_c
    dtype: string
  - name: option_d
    dtype: string
  splits:
  - name: test
    num_bytes: 29088.830825538837
    num_examples: 90
  download_size: 26446
  dataset_size: 29088.830825538837
- config_name: medium
  features:
  - name: answer
    dtype: string
  - name: question
    dtype: string
  - name: difficulty
    dtype: string
  - name: question_type
    dtype: string
  - name: interrogative_type
    dtype: string
  - name: main_medical_field
    dtype: string
  - name: detailed_medical_field
    dtype: string
  - name: option_a
    dtype: string
  - name: option_b
    dtype: string
  - name: option_c
    dtype: string
  - name: option_d
    dtype: string
  splits:
  - name: test
    num_bytes: 849393.8601057341
    num_examples: 2628
  download_size: 389572
  dataset_size: 849393.8601057341
configs:
- config_name: all
  data_files:
  - split: test
    path: all/test-*
- config_name: easy
  data_files:
  - split: test
    path: easy/test-*
- config_name: hard
  data_files:
  - split: test
    path: hard/test-*
- config_name: medium
  data_files:
  - split: test
    path: medium/test-*
---
### **Dataset Card**

## **General Description**
This dataset is a valuable resource created to **evaluate the performance** of large language models (LLMs) in the field of medicine in Vietnamese. It provides a diverse set of questions and answers, categorized by complexity, which helps researchers assess a model's capabilities from a basic to an advanced level.

---

## **Data Source**
The data was compiled and reviewed by a team of experts, including **medical residents** and **master's degree holders in medicine**. The questions are built upon official medical knowledge from reliable documents and textbooks, ensuring high accuracy and trustworthiness. 

---

## **Data Structure and Difficulty Levels**
The dataset is organized into two main categories to provide a comprehensive evaluation of a model's capabilities:

* **Basic Level:** Includes questions about general, common, and fundamental medical knowledge and definitions. These questions do not require complex reasoning, primarily testing the model's ability to recall and retrieve information.
    * **Example:** "What is the normal blood pressure for an adult?" or "What are the functions of Vitamin C?"

* **Advanced Level:** Includes in-depth questions focusing on clinical scenarios and complex medical cases. To answer correctly, the model needs to be able to **reason logically**, synthesize information from multiple sources, and draw conclusions.
    * **Example:** "What are the differential diagnoses for a patient with a high fever and a rash with subcutaneous hemorrhages?" or "Analyze the symptoms to provide initial treatment guidelines for a patient with suspected acute appendicitis."

In total, the dataset contains **nearly 5000 questions**, creating a comprehensive and rich benchmark.

---

## **Intended Use**
This dataset is built **solely for academic and research purposes**.

* **Scientific Research:** Supports researchers in evaluating and improving Vietnamese language models in the medical domain.
* **Education:** Provides students and those in medical education with additional reference materials and practice.

---

## **Disclaimer**
This is a dataset for model evaluation and **is not a substitute for professional medical advice**. Although it has been carefully reviewed, the dataset may **unavoidably contain errors**. Users should note:
* **Do not use** this data for self-diagnosis, treating illnesses, or making any medical decisions.
* All medical information should be **referenced and confirmed** by qualified medical professionals.

Using this dataset implies that you agree to the terms of this disclaimer.