Scholarlink

This chapter introduces a three-level framework for utilising LLMs in Corpus Translation Studies, encompassing input preparation, model-assisted processing, and results visualisation. The system employs XML pre-processing, prompt engineering, and the Translation Annotator desktop app to facilitate interactive analysis of LLM-generated annotations. LLMs are positioned as collaborators that support human interpretation, enabling comparative analysis and iterative refinement. Future work includes fine-tuning with expert annotations, workflow integration for model improvement, and the application of transparency techniques. Thoughtful integration of LLMs can enhance research methodologies in CTS, but calibration to field standards and the needs of researchers is essential.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Empowering Corpus Translation Studies with LLMs

  • 章节

  • 出版模式：

    开放获取

  • 发表日期：

    2026年

查看全文
  • Anna Maria Cipriani,
  • Federico Milana

摘要

This chapter introduces a three-level framework for utilising LLMs in Corpus Translation Studies, encompassing input preparation, model-assisted processing, and results visualisation. The system employs XML pre-processing, prompt engineering, and the Translation Annotator desktop app to facilitate interactive analysis of LLM-generated annotations. LLMs are positioned as collaborators that support human interpretation, enabling comparative analysis and iterative refinement. Future work includes fine-tuning with expert annotations, workflow integration for model improvement, and the application of transparency techniques. Thoughtful integration of LLMs can enhance research methodologies in CTS, but calibration to field standards and the needs of researchers is essential.