A Simple but Effective Approach to Improve Structured Language Model Output for Information Extraction

TL;DR

This paper presents G&O, a simple two-step prompting method that significantly improves large language models' ability to generate structured text for information extraction tasks like NER and RE, with minimal extra effort.

Contribution

The paper introduces G&O, a novel two-step pipeline that separates content generation from structuring, enhancing LLM performance in structured text tasks.

Findings

Significant performance improvements in zero-shot NER and RE tasks.

G&O can be combined with other strategies for further gains.

Minimal additional effort required for implementation.

Abstract

Large language models (LLMs) have demonstrated impressive abilities in generating unstructured natural language according to instructions. However, their performance can be inconsistent when tasked with producing text that adheres to specific structured formats, which is crucial in applications like named entity recognition (NER) or relation extraction (RE). To address this issue, this paper introduces an efficient method, G&O, to enhance their structured text generation capabilities. It breaks the generation into a two-step pipeline: initially, LLMs generate answers in natural language as intermediate responses. Subsequently, LLMs are asked to organize the output into the desired structure, using the intermediate responses as context. G&O effectively separates the generation of content from the structuring process, reducing the pressure of completing two orthogonal tasks simultaneously. Tested on zero-shot NER and RE, the results indicate a significant improvement in LLM performance with minimal additional efforts. This straightforward and adaptable prompting technique can also be combined with other strategies, like self-consistency, to further elevate LLM capabilities in various structured text generation tasks.