[rahimnathwani]

It worked for me on the mobile app.

This Tweet has some sample code at the bottom: https://twitter.com/rez0__/status/1745545813512663203?t=icds...

If I've understood correctly, GPT-4 interprets the code points from the Unicode tag block as corresponding to letters. If so, I'm curious to know whether this is something it learned or whether it was hard-coded.

[goodside]

I can’t imagine it was intentionally added as a feature. It doesn’t work in GPT-3.5 — it seems GPT-4 is unexpectedly smart enough to parse the invisible portion (and confuse it for user instruction) whereas in any other context it’s just steganography that would need to be decoded explicitly.

[dietr1ch]

I'd guess that the tokenizer is just different and handles this in a "better" way.

[goodside]

No, in both tokenizers Unicode tag-block code points like these are converted into bytes (two tokens per character), which is a fallback for code points uncommon enough to not warrant a dedicated token.

[rahimnathwani]

How and why would the tokenizer learn that particular unicode tag was equivalent to a particular letter? I can't imagine there's a lot of text on the internet encoded in this way.

[kevingadd]

maybe it saw them used in their intended way (for flags, etc) and was able to make the association between the flags and their country codes, and then that led to it being able to interpret them as individual letters?

could also be from having been trained on unicode character tables, which contain english descriptions of each code point

[rahimnathwani]

BTW these are the character tables:

https://unicode.org/charts/PDF/UE0000.pdf