3 ms·
They have already streamed the first part of the response before the filtered phrase has even been generated.
by Gigachad 2y ago
They have already streamed the first part of the response before the filtered phrase has even been generated.
- _fzslm 2y agoCould you stream the raw tokens into a server side filter which then streams censored tokens at near real time?