Datasets:
Tasks:
Question Answering
Modalities:
Text
Formats:
json
Languages:
French
Size:
10K - 100K
ArXiv:
License:
Update README.md
Browse files
README.md
CHANGED
|
@@ -9,4 +9,52 @@ tags:
|
|
| 9 |
- biology
|
| 10 |
size_categories:
|
| 11 |
- 10K<n<100K
|
| 12 |
-
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 9 |
- biology
|
| 10 |
size_categories:
|
| 11 |
- 10K<n<100K
|
| 12 |
+
---
|
| 13 |
+
|
| 14 |
+
# MedInjection-FR — Native Subset 🇫🇷
|
| 15 |
+
|
| 16 |
+
## Summary
|
| 17 |
+
|
| 18 |
+
The **Native** component of **MedInjection-FR** comprises **French biomedical instructions and question–answer pairs** natively written in French.
|
| 19 |
+
It forms the core of the dataset and reflects **authentic medical reasoning and linguistic formulations**, sourced from curated corpora such as *FrBMedQA*, *FrenchMedMCQA*, and *MediQAl*.
|
| 20 |
+
|
| 21 |
+
This subset serves as the **high-quality reference supervision** for instruction tuning of large language models (LLMs) in French biomedical contexts.
|
| 22 |
+
|
| 23 |
+
## Motivation
|
| 24 |
+
|
| 25 |
+
Instruction tuning has become essential for adapting LLMs to domain-specific prompts. Yet, in medicine, **native French supervision remains scarce**.
|
| 26 |
+
The Native subset bridges this gap, providing instruction–response data derived directly from **French medical exams, textbooks, and clinical resources**, preserving authentic phrasing, domain structure, and context.
|
| 27 |
+
|
| 28 |
+
## Composition
|
| 29 |
+
|
| 30 |
+
- **Languages:** French (`fr`)
|
| 31 |
+
- **Domain:** Medicine, Clinical reasoning, Biology
|
| 32 |
+
- **Task types:**
|
| 33 |
+
- Multiple-choice (single and multiple answers)
|
| 34 |
+
- Open-ended questions (diagnostic reasoning, treatment, translation-like reformulations)
|
| 35 |
+
- **Structure:**
|
| 36 |
+
Train/
|
| 37 |
+
mcq/
|
| 38 |
+
mcqu/
|
| 39 |
+
oeq/
|
| 40 |
+
Validation/
|
| 41 |
+
mcq/
|
| 42 |
+
mcqu/
|
| 43 |
+
oeq/
|
| 44 |
+
Test/
|
| 45 |
+
mcq/
|
| 46 |
+
mcqu/
|
| 47 |
+
oeq/
|
| 48 |
+
|
| 49 |
+
java
|
| 50 |
+
Copy code
|
| 51 |
+
Each JSON record includes:
|
| 52 |
+
```json
|
| 53 |
+
{
|
| 54 |
+
"instruction": "...",
|
| 55 |
+
"context": "...", // optional
|
| 56 |
+
"options": ["..."], // for MCQ/MCQU
|
| 57 |
+
"answer": "...",
|
| 58 |
+
"origin": "FrBMedQA",
|
| 59 |
+
"type": "mcqu"
|
| 60 |
+
}
|