What Do Language Models Hear? Probing for Auditory Representations in Language Models

Jerry Ngo, Yoon Kim


Abstract
This work explores whether language models encode meaningfully grounded representations of sounds of objects. We learn a linear probe that retrieves the correct text representation of an object given a snippet of audio related to that object, where the sound representation is given by a pretrained audio model. This probe is trained via a contrastive loss that pushes the language representations and sound representations of an object to be close to one another. After training, the probe is tested on its ability to generalize to objects that were not seen during training. Across different language models and audio models, we find that the probe generalization is above chance in many cases, indicating that despite being trained only on raw text, language models encode grounded knowledge of sounds for some objects.
Anthology ID:
2024.acl-long.297
Volume:
Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)
Month:
August
Year:
2024
Address:
Bangkok, Thailand
Editors:
Lun-Wei Ku, Andre Martins, Vivek Srikumar
Venue:
ACL
SIG:
Publisher:
Association for Computational Linguistics
Note:
Pages:
5435–5448
Language:
URL:
https://aclanthology.org/2024.acl-long.297
DOI:
10.18653/v1/2024.acl-long.297
Bibkey:
Cite (ACL):
Jerry Ngo and Yoon Kim. 2024. What Do Language Models Hear? Probing for Auditory Representations in Language Models. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 5435–5448, Bangkok, Thailand. Association for Computational Linguistics.
Cite (Informal):
What Do Language Models Hear? Probing for Auditory Representations in Language Models (Ngo & Kim, ACL 2024)
Copy Citation:
PDF:
https://aclanthology.org/2024.acl-long.297.pdf