README.md

metadata

license: cc-by-4.0
language:
  - de
  - fr
  - it
task_categories:
  - summarization
dataset_info:
  - config_name: de
    features:
      - name: decision_id
        dtype: string
      - name: decision
        dtype: string
      - name: headnote
        dtype: string
      - name: law_area
        dtype: string
      - name: year
        dtype: int64
      - name: url
        dtype: string
    splits:
      - name: train
        num_bytes: 201723622.56204563
        num_examples: 13550
      - name: validation
        num_bytes: 3383819.6266666665
        num_examples: 137
      - name: test
        num_bytes: 5089582.021472393
        num_examples: 207
    download_size: 106800873
    dataset_size: 210197024.2101847
  - config_name: fr
    features:
      - name: decision_id
        dtype: string
      - name: decision
        dtype: string
      - name: headnote
        dtype: string
      - name: law_area
        dtype: string
      - name: year
        dtype: int64
      - name: url
        dtype: string
    splits:
      - name: train
        num_bytes: 81225393.7046879
        num_examples: 5456
      - name: validation
        num_bytes: 1358467.7333333334
        num_examples: 55
      - name: test
        num_bytes: 2630846.745398773
        num_examples: 107
    download_size: 42509320
    dataset_size: 85214708.18342
  - config_name: it
    features:
      - name: decision_id
        dtype: string
      - name: decision
        dtype: string
      - name: headnote
        dtype: string
      - name: law_area
        dtype: string
      - name: year
        dtype: int64
      - name: url
        dtype: string
    splits:
      - name: train
        num_bytes: 13979223.733266482
        num_examples: 939
      - name: validation
        num_bytes: 197595.30666666667
        num_examples: 8
      - name: test
        num_bytes: 295048.23312883434
        num_examples: 12
    download_size: 7426990
    dataset_size: 14471867.273061983
configs:
  - config_name: de
    data_files:
      - split: train
        path: de/train-*
      - split: validation
        path: de/validation-*
      - split: test
        path: de/test-*
  - config_name: fr
    data_files:
      - split: train
        path: fr/train-*
      - split: validation
        path: fr/validation-*
      - split: test
        path: fr/test-*
  - config_name: it
    data_files:
      - split: train
        path: it/train-*
      - split: validation
        path: it/validation-*
      - split: test
        path: it/test-*

SLDS monolingual subsets

Monolingual pairs from the Swiss Landmark Decisions Summarization dataset (ipst/slds, Rolshoven et al., Findings of EMNLP 2025). Each row pairs a Swiss Federal Supreme Court decision with the headnote the court publishes in the same language.

Construction

Each config keeps exactly the rows of the corresponding ipst/slds split where decision_language and headnote_language both equal the config's language, with the original column values unchanged (decision_id, decision, headnote, law_area, year, url). The cross-lingual pairs of the source dataset are omitted.

Subsets and splits

config	train	validation	test
de	13,550	137	207
fr	5,456	55	107
it	939	8	12

License and citation

CC BY 4.0, following the source dataset. Please cite the source paper:

@inproceedings{rolshoven-etal-2025-unlocking,
  title = "Unlocking Legal Knowledge: A Multilingual Dataset for Judicial Summarization in {S}witzerland",
  author = {Rolshoven, Luca and Rasiah, Vishvaksenan and Bose, Srinanda Br{\"u}gger and Hostettler, Sarah and Burkhalter, Lara and St{\"u}rmer, Matthias and Niklaus, Joel},
  booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2025",
  month = nov,
  year = "2025",
}