Create README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,68 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# ClinBench-HPB: A Clinical Benchmark for Evaluating LLMs in Hepato-Pancreato-Biliary Diseases
|
| 2 |
+
|
| 3 |
+
ClinBench-HPB is a comprehensive benchmark dataset specifically designed to evaluate the performance of Large Language Models (LLMs) in the field of hepatobiliary and pancreatic diseases.
|
| 4 |
+
|
| 5 |
+
## Dataset Overview
|
| 6 |
+
|
| 7 |
+
This dataset consists of five carefully constructed subsets covering different data sources and question types:
|
| 8 |
+
|
| 9 |
+
### 1. CNQA (Chinese Questions & Answers)
|
| 10 |
+
- **2000 Chinese multiple-choice questions**
|
| 11 |
+
- Includes single-choice and multiple-choice questions
|
| 12 |
+
- Each question may have **one or more** correct answers
|
| 13 |
+
|
| 14 |
+
### 2. ENQA (English Questions & Answers)
|
| 15 |
+
- **1535 English multiple-choice questions**
|
| 16 |
+
- Each question has **exactly one** correct answer
|
| 17 |
+
|
| 18 |
+
### 3. Journal
|
| 19 |
+
- **120** open diagnostic cases from authoritative medical journals
|
| 20 |
+
|
| 21 |
+
### 4. Website
|
| 22 |
+
- **167** diagnostic cases from professional medical websites and case-sharing platforms
|
| 23 |
+
|
| 24 |
+
### 5. Hospital
|
| 25 |
+
- **50** real-world diagnostic cases from partner hospitals (anonymized)
|
| 26 |
+
|
| 27 |
+
## Key Features
|
| 28 |
+
|
| 29 |
+
- **Comprehensiveness**: Covers multiple subspecialties of HPB diseases
|
| 30 |
+
|
| 31 |
+
- **Diversity**: Includes both multiple-choice questions and real clinical cases
|
| 32 |
+
|
| 33 |
+
- **High Quality**: All data reviewed by medical experts
|
| 34 |
+
|
| 35 |
+
- **Multilingual**: Supports both Chinese and English evaluation
|
| 36 |
+
|
| 37 |
+
## Usage Suggestions
|
| 38 |
+
|
| 39 |
+
This dataset is suitable for:
|
| 40 |
+
|
| 41 |
+
1. Evaluating LLMs' medical knowledge in HPB diseases
|
| 42 |
+
|
| 43 |
+
2. Comparing model performance on medical QA tasks
|
| 44 |
+
|
| 45 |
+
3. Developing medical education applications
|
| 46 |
+
|
| 47 |
+
4. Evaluating clinical decision support systems
|
| 48 |
+
|
| 49 |
+
## Notice
|
| 50 |
+
|
| 51 |
+
This benchmark is for research purposes only and does not constitute medical advice. Always consult healthcare professionals for clinical decisions. This dataset is licensed for **academic research only**. **Commercial use is strictly prohibited**. No modification, distribution, or use in commercial products is permitted without explicit authorization. If you believe this dataset infringes your rights, please contact us via: Email: yuchong.li@connect.polyu.hk
|
| 52 |
+
|
| 53 |
+
|
| 54 |
+
|
| 55 |
+
## Citation
|
| 56 |
+
|
| 57 |
+
If you use ClinBench-HPB in your research, please cite our work:
|
| 58 |
+
|
| 59 |
+
```bibtex
|
| 60 |
+
@misc{li2025clinbenchhpbclinicalbenchmarkevaluating,
|
| 61 |
+
title={ClinBench-HPB: A Clinical Benchmark for Evaluating LLMs in Hepato-Pancreato-Biliary Diseases},
|
| 62 |
+
author={Yuchong Li and Xiaojun Zeng and Chihua Fang and Jian Yang and Fucang Jia and Lei Zhang},
|
| 63 |
+
year={2025},
|
| 64 |
+
eprint={2506.00095},
|
| 65 |
+
archivePrefix={arXiv},
|
| 66 |
+
primaryClass={cs.CY},
|
| 67 |
+
url={https://arxiv.org/abs/2506.00095},
|
| 68 |
+
}
|