Unambiguous English function words — closed-class tokens that never appear in normal prose for any of the 13 non-English target languages.
Chosen so a translator who forgot to translate a paragraph (or who accidentally pasted the EN source into the target file) is caught at 5–10 % density. The set deliberately excludes ambiguous tokens that legitimately appear as loanwords or proper-noun parts in Swedish / German / Dutch (bill, government, parliament, report).
Unambiguous English function words — closed-class tokens that never appear in normal prose for any of the 13 non-English target languages.
Chosen so a translator who forgot to translate a paragraph (or who accidentally pasted the EN source into the target file) is caught at 5–10 % density. The set deliberately excludes ambiguous tokens that legitimately appear as loanwords or proper-noun parts in Swedish / German / Dutch (
bill,government,parliament,report).