3 ms·
Interesting, do you happen to have some quantitative results on this/additional insights/etc? I've interpreted transformer vector similarity as 'likelihood to
by Nowado 3y ago
Interesting, do you happen to have some quantitative results on this/additional insights/etc?
I've interpreted transformer vector similarity as 'likelihood to be followed by the same thing' which is close to word2vec's 'sum of likelihoods of all words to be replaced by the other set' (kinda), but also very different in some contexts.
- BoorishBears 3y agoThere's no simplified definition like that, vectors can even capture logical properties, it's all down to what the model was tuned for: https://www.sbert.net/examples/training/nli/README.html https://www.sbert.net/examples/training/nli/README.html