Swiss4Arg
Swiss4Arg: A Parallel Dataset for Multilingual Argument Mining across Swiss National Languages
A parallel dataset of 53 Swiss voting issues annotated for argument components and relations in German, French, Italian, and Romansh.
How to run
$ npm i -g lettokenburn-cli$ ltb bench info swiss4arg
External harness. No Harbor adapter has been generated for this benchmark yet. `ltb bench info` prints the known links.
Runs happen on your machine with your own subscription; only the score summary is pushed. CLI docs
Scores
2 sourced entriesEvery number links to where it was published. Low-confidence rows are muted and need per-cell verification.
Reproduced by the community
All runsAbout
Swiss4Arg comprises 53 federal voting issues (2019-2025) from official brochures, aligned across four languages. It features fine-grained annotations for argument components (major claims, claims, premises) and relations (support, attack) by native speakers.
The benchmark evaluates multilingual argument mining models, providing baseline results using GPT-5.4 and SwissBERT. It supports cross-lingual comparability and analysis of annotation disagreement in low-resource settings like Romansh.