Abstract
Based on multilingual pre-trained models, cross-lingual transfer with promptlearning has shown promising effectiveness, where soft prompt learned in asource language is transferred to target languages for downstream tasks,particularly in the low-resource scenario. To efficiently transfer soft prompt,we propose a novel framework, Multilingual Prompt Translator (MPT), where amultilingual prompt translator is introduced to properly process crucialknowledge embedded in prompt by changing language knowledge while retainingtask knowledge. Concretely, we first train prompt in source language and employtranslator to translate it into target prompt. Besides, we extend an externalcorpus as auxiliary data, on which an alignment task for predicted answerprobability is designed to convert language knowledge, thereby equipping targetprompt with multilingual knowledge. In few-shot settings on XNLI, MPTdemonstrates superiority over baselines by remarkable improvements. MPT is moreprominent compared with vanilla prompting when transferring to languages quitedistinct from source language.