Create README.md
Browse files
README.md
ADDED
@@ -0,0 +1,52 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
---
|
2 |
+
language:
|
3 |
+
- en
|
4 |
+
tags:
|
5 |
+
- medical
|
6 |
+
---
|
7 |
+
# MMedS-Bench
|
8 |
+
[💻Github Repo](https://github.com/MAGIC-AI4Med/MedS-Ins) [🖨️arXiv Paper](https://arxiv.org/abs/2408.12547)
|
9 |
+
|
10 |
+
The official benchmark for "Towards Evaluating and Building Versatile Large Language Models for Medicine".
|
11 |
+
|
12 |
+
|
13 |
+
|
14 |
+
## Introduction
|
15 |
+
MedS-Bench is a comprehensive benchmark designed to assess the performance of various large language models (LLMs) in clinical settings. It extends beyond traditional multiple-choice questions to include a wider range of medical tasks, providing a robust framework for evaluating LLM capabilities in healthcare.
|
16 |
+
|
17 |
+
The benchmark is structured around 11 high-level clinical task categories, each derived from a collection of 28 existing datasets. These datasets have been reformatted into an instruction-prompted question-answering format, which includes hand-crafted task definitions to guide the LLM in generating responses. The categories included in MedS-Bench are diverse and cover essential aspects of clinical decision-making and data handling:
|
18 |
+
|
19 |
+
- Multi-choice Question Answering: Tests the ability of LLMs to select correct answers from multiple options based on clinical knowledge.
|
20 |
+
- Text Summarization: Assesses the capability to concisely summarize medical texts.
|
21 |
+
- Information Extraction: Evaluates how effectively an LLM can identify and extract relevant information from complex medical documents.
|
22 |
+
- Explanation and Rationale: Requires the model to provide detailed explanations or justifications for clinical decisions or data.
|
23 |
+
- Named Entity Recognition: Focuses on the ability to detect and classify entities within a medical text.
|
24 |
+
- Diagnosis: Tests diagnostic skills, requiring the LLM to identify diseases or conditions from symptoms and case histories.
|
25 |
+
- Treatment Planning: Involves generating appropriate treatment plans based on patient information.
|
26 |
+
- Clinical Outcome Prediction: Assesses the ability to predict patient outcomes based on clinical data.
|
27 |
+
- Text Classification: Involves categorizing text into predefined medical categories.
|
28 |
+
- Fact Verification: Tests the ability to verify the accuracy of medical facts.
|
29 |
+
- Natural Language Inference: Requires deducing logical relationships from medical text.
|
30 |
+
|
31 |
+
Notably, as the evaluation involves commercial models, for example, GPT-4 and Claude 3.5, it is extremely costly to adopt the original large-scale test split. **Therefore, for some benchmarks, we randomly sampling a number of test cases.** The cases used to reeproduce the results in the paper are in [MedS-Bench-SPLIT](https://huggingface.co/datasets/Henrychur/MedS-Bench-SPLIT). For more details, please refer to our paper。
|
32 |
+
|
33 |
+
## Data Format
|
34 |
+
The data format is the same as [MedS-Ins](https://huggingface.co/datasets/Henrychur/MedS-Ins).
|
35 |
+
```bash
|
36 |
+
{
|
37 |
+
"Contributors": [""],
|
38 |
+
"Source": [""],
|
39 |
+
"URL": [""],
|
40 |
+
"Categories": [""],
|
41 |
+
"Reasoning": [""],
|
42 |
+
"Definition": [""],
|
43 |
+
"Input_language": [""],
|
44 |
+
"Output_language": [""],
|
45 |
+
"Instruction_language": [""],
|
46 |
+
"Domains": [""],
|
47 |
+
"Positive Examples": [ { "input": "", "output": "", "explanation": ""} ],
|
48 |
+
"Negative Examples": [ { "input": "", "output": "", "explanation": ""} ],
|
49 |
+
"Instances": [ { "id": "", "input": "", "output": [""]} ],
|
50 |
+
}
|
51 |
+
```
|
52 |
+
|