wow this was hard but fun. apparently i don't understand the structure of word2vec at all
semantle.novalis.org
Conversation
i was thrown off by a bunch of nearby words that had several very distinct meanings, i think some of the similarities were for different meanings and that took me a long time to sort out
1
7
the word2vec manifold must be folded in on itself in weird ways by words with multiple meanings, i was i guess naively expecting it to be a lot "flatter"
3
1
10
oh nice the creator is on twitter. i asked him for a 2-player version
Quote Tweet
Is Wordle too easy for you? Try Semantle, which is based on semantics (well, word2vec) rather than letters. New puzzle every day, as seems to be the convention. semantle.novalis.org
1
1
7
ah apparently it's expensive to generate and store the ranking data for each word, sad times
Quote Tweet
Replying to @QiaochuYuan
Or just store it in base64 and assume the user won't cheat. One problem is that I've only generated the top 1k nearest for ~4k of the 3m tokens in the dataset. I might decide to live without that, or decide to generate a few hundred k more and pay for a bit more disk.