The remarkable ability of Large Language Models (LLMs) to understand and follow instructions has sometimes been limited by their in-context learning (ICL) performance in low-resource languages. To address this, we introduce a novel approach that leverages cross-lingual retrieval-augmented in-context learning (CREA-ICL). By extracting semantically similar prompts from high-resource languages, we aim to improve the zero-shot performance of multilingual pre-trained language models (MPLMs) across diverse tasks. Though our approach yields steady improvements in classification tasks, it faces challenges in generation tasks. Our evaluation offers insights into the performance dynamics of retrieval-augmented in-context learning across both classification and generation domains.
@article{arxiv.2311.06595,
title = {From Classification to Generation: Insights into Crosslingual Retrieval Augmented ICL},
author = {Xiaoqian Li and Ercong Nie and Sheng Liang},
journal= {arXiv preprint arXiv:2311.06595},
year = {2023}
}
Comments
In The Workshop on Instruction Tuning and Instruction Following, held in conjunction with The Conference on NeurIPS 2023, December 2023. arXiv admin note: text overlap with arXiv:2311.00587