How Auditory Knowledge in LLM Backbones Shapes Audio Language Models: A Holistic Evaluation

TL;DR

This paper investigates how large language models encode auditory knowledge through text-only pre-training and how this impacts their performance in audio-related tasks, using various evaluation methods.

Contribution

It provides a comprehensive empirical analysis of auditory knowledge in LLMs and its influence on audio language model performance across multiple evaluation settings.

Findings

Auditory knowledge varies significantly across LLM families.

Text-only performance correlates strongly with audio task performance.

Empirical insights into how LLMs can be effectively used in audio applications.

Abstract

Large language models (LLMs) have been widely used as knowledge backbones of Large Audio Language Models (LALMs), yet how much auditory knowledge they encode through text-only pre-training and how this affects downstream performance remains unclear. We study this gap by comparing different LLMs under two text-only and one audio-grounded setting: (1) direct probing on AKB-2000, a curated benchmark testing the breadth and depth of auditory knowledge; (2) cascade evaluation, where LLMs reason over text descriptions from an audio captioner; and (3) audio-grounded evaluation, where each LLM is fine-tuned into a Large Audio Language Model (LALM) with an audio encoder. Our findings reveal that auditory knowledge varies substantially across families, and text-only results are strongly correlated with audio performance. Our work provides empirical grounding for a comprehensive understanding of LLMs in audio research.