Datasets:
ArXiv:
Add dataset card metadata.
Browse filesAdd languages, pretty name to dataset card metadata.
Also update name for Croatian Wikipedia (hr) in README.
README.md
CHANGED
|
@@ -1,3 +1,57 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
# MegaWika 2
|
| 2 |
|
| 3 |
MegaWika 2 is an improved multilingual text dataset containing a structured view of Wikipedia articles, the web sources they cite, source text quality estimates, article text translations, and additional article enrichments.
|
|
@@ -40,7 +94,7 @@ As in MegaWika 1, MegaWika 2 spans 50 languages, including English, designated b
|
|
| 40 |
- `gu`: Gujarati
|
| 41 |
- `he`: Hebrew
|
| 42 |
- `hi`: Hindi
|
| 43 |
-
- `hr`:
|
| 44 |
- `id`: Indonesian
|
| 45 |
- `it`: Italian
|
| 46 |
- `ja`: Japanese
|
|
@@ -145,4 +199,4 @@ Additional statistics are provided in the metrics files (for example, `en/metric
|
|
| 145 |
MegaWika 2 also introduces improvements to error handling, providing higher coverage across the board.
|
| 146 |
Errors and metadata for source scraping and extraction are included in the data, enabling analysis of sources of missing data and potential biases in the data.
|
| 147 |
|
| 148 |
-
For additional details and analysis of the MegaWika 2.0 dataset and its construction, please see our [whitepaper on ArXiv](https://arxiv.org/abs/2508.03828).
|
|
|
|
| 1 |
+
---
|
| 2 |
+
language:
|
| 3 |
+
- af
|
| 4 |
+
- ar
|
| 5 |
+
- az
|
| 6 |
+
- bn
|
| 7 |
+
- cs
|
| 8 |
+
- de
|
| 9 |
+
- en
|
| 10 |
+
- es
|
| 11 |
+
- et
|
| 12 |
+
- fa
|
| 13 |
+
- fi
|
| 14 |
+
- fr
|
| 15 |
+
- ga
|
| 16 |
+
- gl
|
| 17 |
+
- gu
|
| 18 |
+
- he
|
| 19 |
+
- hi
|
| 20 |
+
- hr
|
| 21 |
+
- id
|
| 22 |
+
- it
|
| 23 |
+
- ja
|
| 24 |
+
- ka
|
| 25 |
+
- kk
|
| 26 |
+
- km
|
| 27 |
+
- ko
|
| 28 |
+
- lt
|
| 29 |
+
- lv
|
| 30 |
+
- mk
|
| 31 |
+
- ml
|
| 32 |
+
- mn
|
| 33 |
+
- mr
|
| 34 |
+
- my
|
| 35 |
+
- ne
|
| 36 |
+
- nl
|
| 37 |
+
- pl
|
| 38 |
+
- ps
|
| 39 |
+
- pt
|
| 40 |
+
- ro
|
| 41 |
+
- ru
|
| 42 |
+
- si
|
| 43 |
+
- sl
|
| 44 |
+
- sv
|
| 45 |
+
- ta
|
| 46 |
+
- th
|
| 47 |
+
- tr
|
| 48 |
+
- uk
|
| 49 |
+
- ur
|
| 50 |
+
- vi
|
| 51 |
+
- xh
|
| 52 |
+
- zh
|
| 53 |
+
pretty_name: MegaWika 2
|
| 54 |
+
---
|
| 55 |
# MegaWika 2
|
| 56 |
|
| 57 |
MegaWika 2 is an improved multilingual text dataset containing a structured view of Wikipedia articles, the web sources they cite, source text quality estimates, article text translations, and additional article enrichments.
|
|
|
|
| 94 |
- `gu`: Gujarati
|
| 95 |
- `he`: Hebrew
|
| 96 |
- `hi`: Hindi
|
| 97 |
+
- `hr`: Croatian
|
| 98 |
- `id`: Indonesian
|
| 99 |
- `it`: Italian
|
| 100 |
- `ja`: Japanese
|
|
|
|
| 199 |
MegaWika 2 also introduces improvements to error handling, providing higher coverage across the board.
|
| 200 |
Errors and metadata for source scraping and extraction are included in the data, enabling analysis of sources of missing data and potential biases in the data.
|
| 201 |
|
| 202 |
+
For additional details and analysis of the MegaWika 2.0 dataset and its construction, please see our [whitepaper on ArXiv](https://arxiv.org/abs/2508.03828).
|