Poor embedding performance using ada for portuguse

[_j]

For single word embeddings in a distinct language, I think you picked the wrong model. ada is best for comparing larger texts and user inputs to texts.

Repositório de Word Embeddings do NILC

The best performance for semantic word embeddings there is reported as glove 600. You can see how to employ the Gensim model and obtain single word vectors here:

Word2vec Tutorial | RARE Technologies