Benchmarking Linguistic Diversity of Large Language Models

TL;DR

This paper introduces a comprehensive framework for evaluating the linguistic diversity of large language models, assessing their lexical, syntactic, and semantic richness compared to human language.

Contribution

It proposes a novel evaluation framework for linguistic diversity and benchmarks several state-of-the-art LLMs across multiple linguistic dimensions.

Findings

LLMs vary significantly in lexical diversity

Syntactic diversity is often limited in current models

Development choices influence linguistic richness of outputs

Abstract

The development and evaluation of Large Language Models (LLMs) has primarily focused on their task-solving capabilities, with recent models even surpassing human performance in some areas. However, this focus often neglects whether machine-generated language matches the human level of diversity, in terms of vocabulary choice, syntactic construction, and expression of meaning, raising questions about whether the fundamentals of language generation have been fully addressed. This paper emphasizes the importance of examining the preservation of human linguistic richness by language models, given the concerning surge in online content produced or aided by LLMs. We propose a comprehensive framework for evaluating LLMs from various linguistic diversity perspectives including lexical, syntactic, and semantic dimensions. Using this framework, we benchmark several state-of-the-art LLMs across all diversity dimensions, and conduct an in-depth case study for syntactic diversity. Finally, we analyze how different development and deployment choices impact the linguistic diversity of LLM outputs.