English

Source-primed Multi-turn Conversation Helps Large Language Models Translate Documents

Computation and Language 2025-03-14 v1

Abstract

LLMs have paved the way for truly simple document-level machine translation, but challenges such as omission errors remain. In this paper, we study a simple method for handling document-level machine translation, by leveraging previous contexts in a multi-turn conversational manner. Specifically, by decomposing documents into segments and iteratively translating them while maintaining previous turns, this method ensures coherent translations without additional training, and can fully re-use the KV cache of previous turns thus minimizing computational overhead. We further propose a `source-primed' method that first provides the whole source document before multi-turn translation. We empirically show this multi-turn method outperforms both translating entire documents in a single turn and translating each segment independently according to multiple automatic metrics in representative LLMs, establishing a strong baseline for document-level translation using LLMs.

Keywords

Cite

@article{arxiv.2503.10494,
  title  = {Source-primed Multi-turn Conversation Helps Large Language Models Translate Documents},
  author = {Hanxu Hu and Jannis Vamvas and Rico Sennrich},
  journal= {arXiv preprint arXiv:2503.10494},
  year   = {2025}
}

Comments

9 pages, 2 figures

R2 v1 2026-06-28T22:19:15.041Z