Do Audio-Visual Large Language Models Really See and Hear?

TL;DR

This study investigates how audio-visual large language models process and fuse audio and visual information, revealing a bias towards visual data that limits audio influence in final text outputs.

Contribution

First mechanistic interpretability analysis of AVLLMs showing modality bias and limited audio integration due to training and fusion layer dynamics.

Findings

AVLLMs encode rich audio semantics at intermediate layers.

Final text outputs are dominated by visual information, often suppressing audio cues.

Training aligns AVLLMs more with vision-language models than with audio supervision.

Abstract

Audio-Visual Large Language Models (AVLLMs) are emerging as unified interfaces to multimodal perception. We present the first mechanistic interpretability study of AVLLMs, analyzing how audio and visual features evolve and fuse through different layers of an AVLLM to produce the final text outputs. We find that although AVLLMs encode rich audio semantics at intermediate layers, these capabilities largely fail to surface in the final text generation when audio conflicts with vision. Probing analyses show that useful latent audio information is present, but deeper fusion layers disproportionately privilege visual representations that tend to suppress audio cues. We further trace this imbalance to training: the AVLLM's audio behavior strongly matches its vision-language base model, indicating limited additional alignment to audio supervision. Our findings reveal a fundamental modality bias in AVLLMs and provide new mechanistic insights into how multimodal LLMs integrate audio and vision.