A Comparative Analysis of Conversational Large Language Models in Knowledge-Based Text Generation

TL;DR

This paper empirically compares conversational large language models in generating natural language from knowledge graph triples, highlighting how prompting and fine-tuning improve their factual accuracy and consistency.

Contribution

It provides a comparative analysis of multiple large language models using benchmark experiments and explores techniques to enhance their knowledge-based text generation.

Findings

Few-shot prompting significantly improves model performance.

Post-processing and fine-tuning enhance factual accuracy.

Smaller models benefit most from targeted tuning.

Abstract

Generating natural language text from graph-structured data is essential for conversational information seeking. Semantic triples derived from knowledge graphs can serve as a valuable source for grounding responses from conversational agents by providing a factual basis for the information they communicate. This is especially relevant in the context of large language models, which offer great potential for conversational interaction but are prone to hallucinating, omitting, or producing conflicting information. In this study, we conduct an empirical analysis of conversational large language models in generating natural language text from semantic triples. We compare four large language models of varying sizes with different prompting techniques. Through a series of benchmark experiments on the WebNLG dataset, we analyze the models' performance and identify the most common issues in the generated predictions. Our findings show that the capabilities of large language models in triple verbalization can be significantly improved through few-shot prompting, post-processing, and efficient fine-tuning techniques, particularly for smaller models that exhibit lower zero-shot performance.