Sparse Autoencoders Make Audio Foundation Models more Explainable
This work provides an incremental improvement in explainability for audio foundation models, primarily benefiting researchers in speech processing and music information retrieval.
The researchers tackled the problem of unclear representations in audio pretrained models by applying Sparse Autoencoders (SAEs) to analyze hidden representations, demonstrating that SAEs retain information about original representations and class labels while enhancing disentanglement of vocal attributes in a singing technique classification case study.
Audio pretrained models are widely employed to solve various tasks in speech processing, sound event detection, or music information retrieval. However, the representations learned by these models are unclear, and their analysis mainly restricts to linear probing of the hidden representations. In this work, we explore the use of Sparse Autoencoders (SAEs) to analyze the hidden representations of pretrained models, focusing on a case study in singing technique classification. We first demonstrate that SAEs retain both information about the original representations and class labels, enabling their internal structure to provide insights into self-supervised learning systems. Furthermore, we show that SAEs enhance the disentanglement of vocal attributes, establishing them as an effective tool for identifying the underlying factors encoded in the representations.