EuroWeb-2512
Multilingual web data filtered for European LLM pretraining.
Open dataset Updated 2026-06-18
What is EuroWeb-2512?
EuroWeb-2512 is a dataset of multilingual web data gathered from multiple sources, processed with standard practices, and classified with EuroFilter-v1.
It supports LLM pretraining focused on European languages.
What you can build with EuroWeb-2512
Pretraining European-focused LLMs
Use the filtered web corpus to continue pretraining or fine-tune models like EuroLLM-22B on high-quality multilingual European data.
Evaluating web data filters
Benchmark new content classifiers by comparing against the EuroFilter-v1 labels already applied to this collection.
Multilingual web research
Analyze language distribution, domain coverage, and quality signals across the collected European web sources.
Load EuroWeb-2512
Python
from datasets import load_dataset
ds = load_dataset("utter-project/EuroWeb-2512")- 1pip install datasets
- 2from datasets import load_dataset
- 3ds = load_dataset('utter-project/EuroWeb-2512')
- 4Inspect available splits and columns
- 5Stream subsets with streaming=True for large-scale use
EuroWeb-2512: pros & cons
Pros
- +Multilingual European web data
- +Already processed with EuroFilter-v1
- +Linked to EuroLLM-22B technical report
- +Directly loadable via Hugging Face datasets
Cons
- –Web data can still contain noise
- –Exact license not stated in summary
- –Size may require substantial storage
Did you find this helpful?
Frequently asked questions
A multilingual web dataset collected from various sources, processed with standard practices and classified using utter-project/EuroFilter-v1.
User reviews
Verified reviews from the community shape this listing's rating.
Loading reviews…
Similar datasets
Other data & analytics options worth comparing.