"Don't Teach Minerva": Guiding LLMs Through Complex Syntax for Faithful Latin Translation with RAG
This addresses the problem of accurate Latin translation for researchers and linguists, offering an open-source alternative to proprietary systems, though it is incremental as it builds on existing methods like RAG and fine-tuning.
The paper tackled the challenge of translating Latin, a morphology-rich, low-resource language, by introducing a draft-based refinement pipeline that uses a fine-tuned NLLB-1.3B model and zero-shot LLMs with RAG, achieving performance statistically comparable to GPT-5 on both in-domain and out-of-domain benchmarks.
Translating a morphology-rich, low-resource language like Latin poses significant challenges. This paper introduces a reproducible draft-based refinement pipeline that elevates open-source Large Language Models (LLMs) to a performance level statistically comparable to top-tier proprietary systems. Our method first uses a fine-tuned NLLB-1.3B model to generate a high-quality, structurally faithful draft. A zero-shot LLM (Llama-3.3 or Qwen3) then polishes this draft, a process that can be further enhanced by augmenting the context with retrieved out-context examples (RAG). We demonstrate the robustness of this approach on two distinct benchmarks: a standard in-domain test set (Rosenthal, 2023) and a new, challenging out-of-domain (OOD) set of 12th-century Latin letters (2025). Our central finding is that this open-source RAG system achieves performance statistically comparable to the GPT-5 baseline, without any task-specific LLM fine-tuning. We release the pipeline, the Chartres OOD set, and evaluation scripts and models to facilitate replicability and further research.