Finding
The prompt, v1, frozen
Using LaTeX and TikZ, draw a Venn diagram of three overlapping circles labelled "Has espresso", "Has milk" and "Served cold". Write each of these drinks in the region where it belongs: short black, flat white, iced latte, iced long black, cold brew, babyccino, milkshake, pour-over.
This is the last of four posts on whether specifying to measure helps model avoid overlapping text and lines. Here, with 10 runs instead of 3 for Sol. I had the plain prompt telling Sol to make a Venn diagram, another prompt to make a Venn and check that the text does not overlap with the lines and another to check its own render to ensure no overlap.
| Labels clear of the lines | Sol, ten runs |
|---|---|
| Plain prompt | 9 of 10 |
| Told the list | 9 of 10 |
| Told to check its own render | 9 of 10 |
Verdict
No indication of difference between prompt, but of course, to measure this statistically, I would need a very large sample. So really, we cannot tell yet.
n = 10 runs a prompt. p = 1.0, Fisher exact, committed before the runs.
The 30 runs cost $1.92.