engCorpus-based Semantic Composition Models for Phrases
engContains five datasets: two for German composition (de-nn-only-comp, de-mixed-comp),
one for English composition (en-comp), one containing semantically annotated German compounds (de-nncom-sem) and
one containing German compounds annotated for lexicalization (de-ulex). Includes also word vectors in four dimensions
for German and English words and compounds.