Better Literary Translation: A Multi-Aspect Data Generation and LLM Training Approach
For literary translation researchers and practitioners, this work provides a data generation and training approach that achieves state-of-the-art results on a specific benchmark, though the method is incremental (combining existing techniques).
Literary translation suffers from data scarcity and the need to balance fluency with literary effect. The authors propose a multi-aspect iterative refinement framework generating high-quality translation data, achieving 8.65 CEA100 improvement over ground truth for SFT and 1.51 additional gain via GRPO, resulting in models competitive with Claude Sonnet 4.5.
Literary translation poses unique challenges due to the scarcity of high-quality annotated data and the need to balance expression fluency with literary effect. We present a multi-aspect iterative refinement framework that generates high-quality translation references and preference data through specialized LLM translators, each targeting a distinct quality dimension. We leverage the generated data for supervised fine-tuning and reinforcement learning. Experiments show that our generated references outperform the original ground truth for SFT by 8.65 CEA100 points. For reinforcement learning, we find that DPO leads to performance degradation in this setting, while leveraging an explicit reward model for GRPO yields an additional 1.51 point improvement. We attribute this to the stability of two-stage training and GRPO's online exploration capability. Our resulting models, LitMT-8B and LitMT-14B, achieve 67.25 and 69.07 CEA100 respectively on the MetaphorTrans English-to-Chinese literary translation benchmark, competitive with Claude Sonnet 4.5 at 68.43, and demonstrate strong generalization to out-of-domain literary work (i.e., O. Henry).