3 ms·
Is this handling Unicode correctly? Seems like a lot of even Latin alphabets are getting mangled.
by numpad0 1y ago
Is this handling Unicode correctly? Seems like a lot of even Latin alphabets are getting mangled.
- int_19h 1y agoIt looks like it's not handling UTF-8 at all and displaying it as if it were Latin-1
- mkl 1y agoI don't think it's actually UTF-8. The data is at https://melonmars.github.io/LatentExplorer/embeddings_2d.json https://melonmars.github.io/LatentExplorer/embeddings_2d.jso... and contains things like "\u00e0\u00a7\u012d\u00e0\u00a6\u013e" with some characters > 0xff (but none above 0x0143, weirdly).