Datasets:
fix: drop questions with empty candidates
Browse files- README.md +3 -3
- data/eu/health.jsonl +0 -0
README.md
CHANGED
|
@@ -59,15 +59,15 @@ This new version introduces several key improvements over the original EusExams
|
|
| 59 |
- **New Data**: Added the 2026 health exams in Spanish, which accounts for the larger discrepancy in size between the `es` and `eu` subsets in the `health` category.
|
| 60 |
|
| 61 |
## Dataset Statistics
|
| 62 |
-
The dataset contains a total of **27,
|
| 63 |
|
| 64 |
| Domain | Spanish (`es`) | Basque (`eu`) |
|
| 65 |
| :--- | :--- | :--- |
|
| 66 |
| **Council** | 928 | 895 |
|
| 67 |
| **Government** | 2,530 | 2,531 |
|
| 68 |
-
| **Health** | 11,387 | 6,
|
| 69 |
| **University** | 1,459 | 1,361 |
|
| 70 |
-
| **Total** | **16,304** | **
|
| 71 |
|
| 72 |
- **Curated by:** HiTZ Research Center & IXA Research group (University of the Basque Country UPV/EHU)
|
| 73 |
- **Language(s) (NLP):** Basque (eu)
|
|
|
|
| 59 |
- **New Data**: Added the 2026 health exams in Spanish, which accounts for the larger discrepancy in size between the `es` and `eu` subsets in the `health` category.
|
| 60 |
|
| 61 |
## Dataset Statistics
|
| 62 |
+
The dataset contains a total of **27,096** unique questions, distributed across the two languages and domains as follows:
|
| 63 |
|
| 64 |
| Domain | Spanish (`es`) | Basque (`eu`) |
|
| 65 |
| :--- | :--- | :--- |
|
| 66 |
| **Council** | 928 | 895 |
|
| 67 |
| **Government** | 2,530 | 2,531 |
|
| 68 |
+
| **Health** | 11,387 | 6,005 |
|
| 69 |
| **University** | 1,459 | 1,361 |
|
| 70 |
+
| **Total** | **16,304** | **10,792** |
|
| 71 |
|
| 72 |
- **Curated by:** HiTZ Research Center & IXA Research group (University of the Basque Country UPV/EHU)
|
| 73 |
- **Language(s) (NLP):** Basque (eu)
|
data/eu/health.jsonl
CHANGED
|
The diff for this file is too large to render.
See raw diff
|
|
|