# Explaining Speaker and Spoof Embeddings via Probing

arXiv: https://arxiv.org/abs/2412.18191 · HTML: https://mlsanity.com/paper/2412.18191/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 24 Dec 2024.

## Built on by (most cited first, 2 of 2)

- [Sparse deepfake detection promotes better disentanglement](https://mlsanity.com/paper/2510.05696/) (2025)
- [From Scores to Evidence: Auditable Decisions Can Improve Speech Deepfake Detection](https://mlsanity.com/paper/2609.08899/) (2026)

## Abstract (arXiv)

This study investigates the explainability of embedding representations, specifically those used in modern audio spoofing detection systems based on deep neural networks, known as spoof embeddings. Building on established work in speaker embedding explainability, we examine how well these spoof embeddings capture speaker-related information. We train simple neural classifiers using either speaker or spoof embeddings as input, with speaker-related attributes as target labels. These attributes are categorized into two groups: metadata-based traits (e.g., gender, age) and acoustic traits (e.g., fundamental frequency, speaking rate). Our experiments on the ASVspoof 2019 LA evaluation set demonstrate that spoof embeddings preserve several key traits, including gender, speaking rate, F0, and duration. Further analysis of gender and speaking rate indicates that the spoofing detector partially preserves these traits, potentially to ensure the decision process remains robust against them.