Reproduce “Emotion Profiling in LLM-Based Literary Translation: Systematic Shifts Across MT and Post-Editing” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Antonio Castaldo, Johanna Monti, Sheila Castilho
This study explores how Large Language Model (LLM) translations introduce distinct emotional profiles in literary texts and how subsequent post-editing (PE) modifies these profiles. It identifies systematic, model-specific emotional shifts and demonstrates that PE generally improves emotional density and intensity towards human translation standards, though its effectiveness in preserving specific emotional peaks varies depending on the initial machine translation quality.
Key insights
- Large Language Models introduce systematic, model-specific emotional "fingerprints" in translated texts, manifesting as distinct patterns in emotional distribution and intensity.
- Post-editing generally improves the emotional density and intensity of machine translations, bringing them closer to human translation standards, but its effectiveness is highly dependent on the initial quality of the MT output.
- The quality of the MT output influences translator engagement; higher-quality MT might lead to widespread minor adjustments, while mid-quality MT may necessitate more targeted revisions to recover specific emotional peaks.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.