Upload README.md with huggingface_hub
Browse files
README.md
CHANGED
|
@@ -35,12 +35,6 @@ dataset_info:
|
|
| 35 |
|
| 36 |
A diagnostic benchmark of **100 condition-dependent biomedical questions** where the correct answer changes based on patient-specific factors such as comorbidities, allergies, pregnancy status, and contraindications.
|
| 37 |
|
| 38 |
-
## Paper
|
| 39 |
-
|
| 40 |
-
> **Condition-Gated Reasoning for Context-Dependent Biomedical Question Answering**
|
| 41 |
-
> Jash Rajesh Parekh, Wonbin Kweon, Joey Chan, Rezarta Islamaj, Robert Leaman, Pengcheng Jiang, Chih-Hsuan Wei, Zhizheng Wang, Zhiyong Lu, Jiawei Han
|
| 42 |
-
> [arXiv:2602.17911](https://arxiv.org/abs/2602.17911)
|
| 43 |
-
|
| 44 |
## Dataset Description
|
| 45 |
|
| 46 |
Real-world clinical reasoning is inherently conditional: treatment decisions depend on patient-specific factors such as comorbidities, drug allergies, pregnancy, and organ function. Existing biomedical QA benchmarks largely ignore this conditionality, testing only context-free factual recall.
|
|
@@ -56,12 +50,6 @@ Real-world clinical reasoning is inherently conditional: treatment decisions dep
|
|
| 56 |
- **Answer length**: Typically 1–5 words
|
| 57 |
- **Split**: Test only (diagnostic benchmark)
|
| 58 |
|
| 59 |
-
### Example
|
| 60 |
-
|
| 61 |
-
| id | wiki1 | wiki2 | q | a |
|
| 62 |
-
|----|-------|-------|---|---|
|
| 63 |
-
| 1 | [Lyme disease](https://en.wikipedia.org/wiki/Lyme_disease) | [Doxycycline](https://en.wikipedia.org/wiki/Doxycycline) | What antibiotic is recommended for early localized Lyme disease in a pregnant patient with no penicillin allergy, who cannot take cephalosporins due to a severe allergy? | Amoxicillin |
|
| 64 |
-
|
| 65 |
## Dataset Format
|
| 66 |
|
| 67 |
A single CSV file (`condmedqa.csv`) with the following columns:
|
|
|
|
| 35 |
|
| 36 |
A diagnostic benchmark of **100 condition-dependent biomedical questions** where the correct answer changes based on patient-specific factors such as comorbidities, allergies, pregnancy status, and contraindications.
|
| 37 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 38 |
## Dataset Description
|
| 39 |
|
| 40 |
Real-world clinical reasoning is inherently conditional: treatment decisions depend on patient-specific factors such as comorbidities, drug allergies, pregnancy, and organ function. Existing biomedical QA benchmarks largely ignore this conditionality, testing only context-free factual recall.
|
|
|
|
| 50 |
- **Answer length**: Typically 1–5 words
|
| 51 |
- **Split**: Test only (diagnostic benchmark)
|
| 52 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 53 |
## Dataset Format
|
| 54 |
|
| 55 |
A single CSV file (`condmedqa.csv`) with the following columns:
|