An interesting blog post on 視覴, a ghost word in Japanese (or at least, #CJK) created by ChatGPT, and propagated by content farms presumably using #ChatGPT.
https://okumuralab.org/~okumura/misc/230611.html
It seems like the cause has to do with UTF-8 encoding and imprecise tokenization. #UTF8 #NLProc #GlitchInTheMatrix
In
コメントを残す