# Probing self-attention in self-supervised speech models for cross-linguistic differences

arXiv: https://arxiv.org/abs/2409.03115 · HTML: https://mlsanity.com/paper/2409.03115/

## Key facts

- Cited by 3 papers in the map (data through 14 Oct 2026).
- Builds on 7 of its 10 references that are in the map.
- First posted on arXiv 4 Sep 2024.

## Builds on (most cited first, 7 of 7)

- [wav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representations](https://mlsanity.com/paper/2006.11477/) (2020)
- [Revealing the Dark Secrets of BERT](https://mlsanity.com/paper/1908.08593/) (2019)
- [Self-Supervised Speech Representation Learning: A Review](https://mlsanity.com/paper/2205.10643/) (2022)
- [TERA: Self-Supervised Learning of Transformer Encoder Representation for Speech](https://mlsanity.com/paper/2007.06028/) (2020)
- [Comparative layer-wise analysis of self-supervised speech models](https://mlsanity.com/paper/2211.03929/) (2022)
- [SparseBERT: Rethinking the Importance Analysis in Self-attention](https://mlsanity.com/paper/2102.12871/) (2021)
- [Probing self-supervised speech models for phonetic and phonemic information: a case study in aspiration](https://mlsanity.com/paper/2306.06232/) (2023)

## Built on by (most cited first, 3 of 3)

- [Understanding Multilingualism in Mixture-of-Experts LLMs: Routing Mechanism, Expert Specialization, and Layerwise Steering](https://mlsanity.com/paper/2601.14050/) (2026)
- [Indic-TunedLens: Interpreting Multilingual Models in Indian Languages](https://mlsanity.com/paper/2602.15038/) (2026)
- [Automated Attention Pattern Discovery at Scale in Large Language Models](https://mlsanity.com/paper/2604.03764/) (2026)

## Abstract (arXiv)

Speech models have gained traction thanks to increase in accuracy from novel transformer architectures. While this impressive increase in performance across automatic speech recognition (ASR) benchmarks is noteworthy, there is still much that is unknown about the use of attention mechanisms for speech-related tasks. For example, while it is assumed that these models are learning language-independent (i.e., universal) speech representations, there has not yet been an in-depth exploration of what it would mean for the models to be language-independent. In the current paper, we explore this question within the realm of self-attention mechanisms of one small self-supervised speech transformer model (TERA). We find that even with a small model, the attention heads learned are diverse ranging from almost entirely diagonal to almost entirely global regardless of the training language. We highlight some notable differences in attention patterns between Turkish and English and demonstrate that the models do learn important phonological information during pretraining. We also present a head ablation study which shows that models across languages primarily rely on diagonal heads to classify phonemes.