@inProceedings{morger-2024-swediagnostics-341148, title = {SweDiagnostics: A Diagnostics Natural Language Inference Dataset for Swedish}, abstract = {This paper presents SweDiagnostics, a natural language inference dataset for Swedish based on the GLUE Diagnostic dataset. It is the largest, manually corrected NLI dataset in Swedish to date and can be used to evaluate models on NLI in Swedish as well as estimate English-Swedish language transfer capabilities. We present the dataset, the methodology used for translation, compare existing implementations and discuss limitations of the dataset, in particular those related to translationese.}, booktitle = {17th Workshop on Building and Using Comparable Corpora, BUCC 2024 at LREC-COLING 2024 - Proceedings}, author = {Morger, Felix}, year = {2024}, ISBN = {9782493814319}, }