Mapping for Eng1000?

[csinva]

Hello, thanks for releasing this great dataset!

Just wondering: is there a list/dictionary somewhere that tells us which element of the eng1000 vector corresponds to what semantic feature? I can't seem to find it in the repo.

Thanks,
Chandan

[alexhuth]

Hey Chandan — The code for generating the Eng1000 embedding space is here: https://github.com/HuthLab/english1000 and in particular the set of target/basis words is here: https://github.com/HuthLab/english1000/blob/master/textcore/word_sets.py#L428

[csinva]

Sweet, thank you!