I am using a translation platform to translate a long document (20,000 words).
The in-built machine translation works quite well for this type of document, translation engine being DeepL.
However, after giving it a try, it looks like ChatGPT 4 is providing slightly better suggestions.
In order for me to use ChatGPT as suggestions, I can first export (from my translation platform) the source (that is, the text in its original language, including the formatting tags) to Word.
This export is a Word file with a number of tables, but only two columns of these tables are relevant (untranslated word in a column, translation to be added in the next column).
ChatGPT is a bit funny when you try to prompt it to translate column 1, line 1 and paste the translation in column 1, line 1 - and so on.
However ChatGPT works much better when I provide it with a table of a single column and ask to translate it directly. This means i have to create a “intermediate” Word file containaing only the source text table.
I am now looking for an automation that would allow me to do this and minimise the number of manual copy/paste across documents. I could probably use Google docs as an intermediary format if it came to it.
Thanks