Update README.md
Browse files
README.md
CHANGED
@@ -4,4 +4,29 @@ tags:
|
|
4 |
- medical
|
5 |
size_categories:
|
6 |
- n<1K
|
7 |
-
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
4 |
- medical
|
5 |
size_categories:
|
6 |
- n<1K
|
7 |
+
---
|
8 |
+
# A Comparative Study of Open-Source Large Language Models
|
9 |
+
|
10 |
+
## Dataset Overview
|
11 |
+
|
12 |
+
Welcome to the dataset repository for our paper, "A Comparative Study of Open-Source Large Language Models, GPT-4 and Claude 2: Multiple-Choice Test Taking in Nephrology." The preprint of the paper can be accessed [here](https://arxiv.org/abs/2308.04709).
|
13 |
+
|
14 |
+
## Files
|
15 |
+
|
16 |
+
This repository contains two key files:
|
17 |
+
|
18 |
+
1. **NEJM_All_Questions_And_Answers.csv**: This file includes all the questions and corresponding answers used in the study.
|
19 |
+
|
20 |
+
2. **Ground_Truth_Answers.csv**: This file provides ground truth explanations associated with the questions in the main dataset.
|
21 |
+
|
22 |
+
## Usage
|
23 |
+
|
24 |
+
To utilize this dataset for your research or experimentation:
|
25 |
+
|
26 |
+
1. **Download**: Obtain the dataset files from this repository.
|
27 |
+
2. **Load**: Import the dataset into your preferred data analysis or machine learning environment.
|
28 |
+
3. **Explore**: Investigate the questions, answers, and ground truth explanations for your specific use case.
|
29 |
+
|
30 |
+
##Paper
|
31 |
+
Our paper is accepted to NEJM-AI. For now please read the pre-print at the link: https://arxiv.org/abs/2308.04709
|
32 |
+
|