Decomposed Prompting: Probing Multilingual Linguistic Structure Knowledge in Large Language Models

TL;DR

This paper introduces a decomposed prompting method for sequence labeling in multilingual LLMs, improving accuracy and efficiency in zero- and few-shot settings by generating token-specific prompts.

Contribution

The paper proposes a novel decomposed prompting approach that enhances multilingual linguistic structure probing in LLMs, outperforming traditional iterative prompting methods.

Findings

Decomposed prompting surpasses iterative prompting in efficacy.

It improves efficiency in zero- and few-shot scenarios.

Multilingual prompting reveals insights into linguistic knowledge transfer.

Abstract

Probing the multilingual knowledge of linguistic structure in LLMs, often characterized as sequence labeling, faces challenges with maintaining output templates in current text-to-text prompting strategies. To solve this, we introduce a decomposed prompting approach for sequence labeling tasks. Diverging from the single text-to-text prompt, our prompt method generates for each token of the input sentence an individual prompt which asks for its linguistic label. We test our method on the Universal Dependencies part-of-speech tagging dataset for 38 languages, using both English-centric and multilingual LLMs. Our findings show that decomposed prompting surpasses the iterative prompting baseline in efficacy and efficiency under zero- and few-shot settings. Moreover, our analysis of multilingual performance of English-centric LLMs yields insights into the transferability of linguistic knowledge via multilingual prompting.