Follow-up to #22, from a screenshot of the comma version.
The comma sits at the baseline, so between two raised digits it drops away from them and reads as a stray mark on the line. A space separates them without introducing a character that lives at the wrong height.
[1][2] and [1, 2] both still arrive as two raised digits, still in the order the model wrote them, and a run containing a number we were never shown is still left exactly as written.
Worth noting from the same screenshot: the model writes its citations descending, so [2][1] is what it actually sends for a sentence resting on both. The order is deliberately preserved rather than sorted, so the two sources read in the order the model ranked them. Sorting ascending would be a one-line change if that is preferred instead.
Follow-up to #22, from a screenshot of the comma version.
The comma sits at the baseline, so between two raised digits it drops away from them and reads as a stray mark on the line. A space separates them without introducing a character that lives at the wrong height.
`[1][2]` and `[1, 2]` both still arrive as two raised digits, still in the order the model wrote them, and a run containing a number we were never shown is still left exactly as written.
Worth noting from the same screenshot: the model writes its citations descending, so `[2][1]` is what it actually sends for a sentence resting on both. The order is deliberately preserved rather than sorted, so the two sources read in the order the model ranked them. Sorting ascending would be a one-line change if that is preferred instead.
The comma sits at the baseline, so between two raised digits it drops away from
them and reads as a stray mark on the line. A space separates them without
introducing a character that lives at the wrong height.
[1][2] and [1, 2] both still arrive as two raised digits, still in the order the
model wrote them, and a run containing a number we were never shown is still
left exactly as written.
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
Follow-up to #22, from a screenshot of the comma version.
The comma sits at the baseline, so between two raised digits it drops away from them and reads as a stray mark on the line. A space separates them without introducing a character that lives at the wrong height.
[1][2]and[1, 2]both still arrive as two raised digits, still in the order the model wrote them, and a run containing a number we were never shown is still left exactly as written.Worth noting from the same screenshot: the model writes its citations descending, so
[2][1]is what it actually sends for a sentence resting on both. The order is deliberately preserved rather than sorted, so the two sources read in the order the model ranked them. Sorting ascending would be a one-line change if that is preferred instead.