Prompting LLMs: Length Control for Isometric Machine Translation
By: Dávid Javorský, Ondřej Bojar, François Yvon
Potential Business Impact:
Makes computer translations shorter and better.
In this study, we explore the effectiveness of isometric machine translation across multiple language pairs (En$\to$De, En$\to$Fr, and En$\to$Es) under the conditions of the IWSLT Isometric Shared Task 2022. Using eight open-source large language models (LLMs) of varying sizes, we investigate how different prompting strategies, varying numbers of few-shot examples, and demonstration selection influence translation quality and length control. We discover that the phrasing of instructions, when aligned with the properties of the provided demonstrations, plays a crucial role in controlling the output length. Our experiments show that LLMs tend to produce shorter translations only when presented with extreme examples, while isometric demonstrations often lead to the models disregarding length constraints. While few-shot prompting generally enhances translation quality, further improvements are marginal across 5, 10, and 20-shot settings. Finally, considering multiple outputs allows to notably improve overall tradeoff between the length and quality, yielding state-of-the-art performance for some language pairs.
Similar Papers
Same evaluation, more tokens: On the effect of input length for machine translation evaluation using Large Language Models
Computation and Language
Helps computers judge long translations better.
When Many-Shot Prompting Fails: An Empirical Study of LLM Code Translation
Software Engineering
Few examples work best for translating computer code.
Plan-and-Write: Structure-Guided Length Control for LLMs without Model Retraining
Computation and Language
Makes AI write exactly the right amount of words.