TIM: Teaching Large Language Models to Translate with Comparison
Preprint
Abstract
Open-sourced large language models (LLMs) have demonstrated remarkable efficacy in various tasks with instruction tuning. However, these models can sometimes struggle with tasks that require more specialized knowledge such as translation. One possible reason for such deficiency is that instruction tuning aims to generate fluent and coherent text that continues from a given instruction without being constrained by any task-specific requirements. Moreover, it can be more challenging for tuning smaller LLMs with lower-quality training data. To address this issue, we propose a novel framework using examples in comparison to teach LLMs to learn translation. Our approach involves presenting the model with examples of correct and incorrect translations and using a preference loss to guide the model's learning. We evaluate our method on WMT2022 test sets and show that it outperforms existing methods. Our findings offer a new perspective on fine-tuning LLMs for translation tasks and provide a promising solution for generating high-quality translations. Please refer to Github for more details: https://github.com/lemon0830/TIM.Keywords
All Related Versions
This publication has 11 references indexed in Scilit:
- Chain-of-Dictionary Prompting Elicits Translation in Large Language Models2024
- Multilingual Machine Translation with Large Language Models: Empirical Results and Analysis2024
- ParroT: Translating during Chat using Large Language Models tuned with Human Translation and Feedback2023
- In-context Examples Selection for Machine Translation2023
- Self-Instruct: Aligning Language Models with Self-Generated Instructions2023
- UniTE: Unified Translation Evaluation2022
- Non-Parametric Unsupervised Domain Adaptation for Neural Machine Translation2021
- Detecting Hallucinated Content in Conditional Neural Sequence Generation2021
- BLEURT: Learning Robust Metrics for Text Generation2020
- chrF: character n-gram F-score for automatic MT evaluation2015