Optimising ChatGPT for creativity in literary translation: A case study from English into Dutch, Chinese, Catalan and Spanish

This study examines the variability of Chat-GPT machine translation (MT) outputs across six different configurations in four languages,with a focus on creativity in a literary text. We evaluate GPT translations in different text granularity levels, temperature settings and prompting strategies with a Creativity Score formula. We found that prompting ChatGPT with a minimal instruction yields the best creative translations, with "Translate the following text into [TG] creatively" at the temperature of 1.0 outperforming other configurations and DeepL in Spanish, Dutch, and Chinese. Nonetheless, ChatGPT consistently underperforms compared to human translation (HT).

@article{du2025_2504.18221,
  title={ Optimising ChatGPT for creativity in literary translation: A case study from English into Dutch, Chinese, Catalan and Spanish },
  author={ Shuxiang Du and Ana Guerberof Arenas and Antonio Toral and Kyo Gerrits and Josep Marco Borillo },
  journal={arXiv preprint arXiv:2504.18221},
  year={ 2025 }
}