Tag
This study investigates how prompt language and translation-theory-driven prompt design affect the quality of Spanish-Chinese journalistic translations produced by GPT-5.2, finding that theory-driven prompts improve style under expert evaluation despite automated metrics favoring baseline prompts.