eurlex
Datasets
All datasets matching “eurlex”super_eurlexSuper-EURLEX dataset containing legal documents from multiple languages.
The datasets are build/scrapped from the EURLEX Website [https://eur-lex.europa.eu/homepage.html]
With one split per language and sector, because the available features (metadata) differs for each
sector. Therefore, each sample contains the content of a full legal document in up to 3 different
formats. Those are raw HTML and cleaned HTML (if the HTML format was available on the EURLEX website
during the scrapping process) and cleaned text.
The cleaned text should be available for each sample and was extracted from HTML or PDF.
'Cleaned' HTML stands here for minor cleaning that was done to preserve to a large extent the necessary
HTML information like table structures while removing unnecessary complexity which was introduced to the
original documents due to actions like writing each sentence into a new object.
Additionally, each sample contains metadata which was scrapped on the fly, this implies the following
2 things. First, not every sector contains the same metadata. Second, most metadata might be
irrelevant for most use cases.
In our minds the most interesting metadata is the celex-id which is used to identify the legal
document at hand, but also contains a lot of information about the document
see [https://eur-lex.europa.eu/content/tools/eur-lex-celex-infographic-A3.pdf] as well as eurovoc-
concepts, which are labels that define the content of the documents.
Eurovoc-Concepts are, for example, only available for the sectors 1, 2, 3, 4, 5, 6, 9, C, and E.
The Naming of most metadata is kept like it was on the eurlex website, except for converting
it to lower case and replacing whitespaces with '_'.multi_eurlexMultiEURLEX comprises 65k EU laws in 23 official EU languages (some low-ish resource).
Each EU law has been annotated with EUROVOC concepts (labels) by the Publication Office of EU.
As with the English EURLEX, the goal is to predict the relevant EUROVOC concepts (labels);
this is multi-label classification task (given the text, predict multiple labels).multi_eurlexMultiEURLEX comprises 65k EU laws in 23 official EU languages (some low-ish resource).
Each EU law has been annotated with EUROVOC concepts (labels) by the Publication Office of EU.
As with the English EURLEX, the goal is to predict the relevant EUROVOC concepts (labels);
this is multi-label classification task (given the text, predict multiple labels).eur-lex-sumThe EUR-Lex-Sum dataset is a multilingual resource intended for text summarization in the legal domain.
It is based on human-written summaries of legal acts issued by the European Union.
It distinguishes itself by introducing a smaller set of high-quality human-written samples,
each of which have much longer references (and summaries!) than comparable datasets.
Additionally, the underlying legal acts provide a challenging domain-specific application to legal texts,
which are so far underrepresented in non-English languages.
For each legal act, the sample can be available in up to 24 languages
(the officially recognized languages in the European Union);
the validation and test samples consist entirely of samples available in all languages,
and are aligned across all languages at the paragraph level.eurlex-multilingual
MultiEURLEXMultilabelClassification
An MTEB dataset
Massive Text Embedding Benchmark
EU laws in 23 EU languages containing annotated labels for 21 EUROVOC concepts.
Task category
t2c
Domains
Legal, Government, Written
Referencehttps://huggingface.co/datasets/coastalcph/multi_eurlex
How to evaluate on this task
You can evaluate an embedding model on this dataset using the following code:
import mteb
task =… See the full description on the dataset page: https://huggingface.co/datasets/mteb/eurlex-multilingual.eurlex_resources
Dataset Card for EurlexResources: A Corpus Covering the Largest EURLEX Resources
Dataset Summary
This dataset contains large text resources (~179GB in total) from EURLEX that can be used for pretraining language models.
Use the dataset like this:
from datasets import load_dataset
config = "de_caselaw" # {lang}_{resource}
dataset = load_dataset("joelito/eurlex_resources", config, split='train', streaming=True)
Supported Tasks and Leaderboards
The dataset… See the full description on the dataset page: https://huggingface.co/datasets/joelniklaus/eurlex_resources.