back
7 comments
Two users reported the demo was having issues, but tested prior to sharing and just now and works fine for me. Only error I get is when I exceed the token limit, which is basically input limit; for me it appears to be roughly 200ish tokens.

Token counter is linked to in the demo, but for ease of reference here’s direct link:

https://beta.openai.com/tokenizer

Feel free to share the exact inputs and error messages and I will attempt to replicate them.

EDIT: For fun, here’s an example I just pulled from the the linked demo; my input starts with “rephrase following” and ends with “option two:“

__________

rephrase following sentence in Chinese: what is your favorite color?

option one: 我喜欢哪种颜色?

option two: 我的最爱是哪一种颜色?

In a way, the second option is more polite. However, it’s still not really correct to use ‘favorite’. 我喜欢 means I like or prefer something and it doesn’t necessarily have to be my favorite thing.

As for the first option, I don’t think there is such a word.

I believe that the most common way to ask this question would be 我喜欢什么颜色?or 我喜欢某种�色?

Demo is producing gibberish.
Demo is throwing errors
this isn't new, is it? am I missing something?
Correct, it is not new, it was released by EleutherAI back in February 2022.
this isn't new, is it?