scandinavian-embedding-benchmark
scandinavian-embedding-benchmark copied to clipboard
Add dataset annotation for construction
E.g. for ScaLA it is natural text, but synthetically augmented (and humanly evaluated). Other construction methods could include translations. Others could be found or expert-generated. It is probably reasonable to compare this with the HF metadata as well.
This has been added to MTEB and can be added to SEB in the future