Break pseudo token#422
Conversation
a966b4a to
3124fa0
Compare
|
So THAT'S how you deal with color bleed! Good to know, and do want! |
|
That 'B' stuff on the regex took me a bit too long to understand, so a comment would've been appreciated 🙂 Apart from that, the code looks good to me, and worked as expected on a few tests. |
|
Actually I should refactor this code to properly adapt it to the master branch. Right now it's a bit messy. I'll still add the original author as coauthor. |
3124fa0 to
c35baca
Compare
|
I've finally remembered that I had this PR pending. I've just upstreamed it and it seems to work nicely. |
| // (abc:3.12) - increases attention to abc by a multiplier of 3.12 | ||
| // [abc] - decreases attention to abc by a multiplier of 1.1 | ||
| // BREAK - separates the prompt into conceptually distinct parts for sequential processing | ||
| // B - internal helper pattern; prevents 'B' in 'BREAK' from being consumed as normal text |
There was a problem hiding this comment.
I think this is enough to make the regex a little less magical. Thanks!
Co-authored-by: Urs Ganse <urs.ganse@helsinki.fi>
65a929d to
7604e9f
Compare
|
Thank you for your contribution. |
I've noticed that #285 hasn't been updated for a while so I've upstreamed the PR and fixed the regex to correctly identify the BREAK token without deleting the white spaces.
@ursg I've still marked you as the author as most of the work was yours, if you still want to update your PR instead I'll close this.