Metraipolis › archive › dc_5bf0571e01cd

From Isolated Feature to Orbits: Discovering Music Concepts via Multi-SAE Alignment — arXiv:2610.01864

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.01864
License (as archived)
CC0-1.0
Archived
2026-10-06T12:15:08.541Z
SHA-256 of the archived text
13a0bca92b1ffe8f480dfe0e2b165c2f32d72efb5d8173c1962475071133316b
Stable URL of this page
https://metraipolis.com/documento/dc_5bf0571e01cd

How to cite

From Isolated Feature to Orbits: Discovering Music Concepts via Multi-SAE Alignment — arXiv:2610.01864 — arXiv (en), https://arxiv.org/abs/2610.01864 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_5bf0571e01cd (Metraipolis archive dc_5bf0571e01cd, sha256 13a0bca92b1ffe8f480dfe0e2b165c2f32d72efb5d8173c1962475071133316b)
BibTeX
@misc{dc5bf0571e01cd,
  title = {From Isolated Feature to Orbits: Discovering Music Concepts via Multi-SAE Alignment — arXiv:2610.01864},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.01864},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_5bf0571e01cd; sha256 13a0bca92b1ffe8f480dfe0e2b165c2f32d72efb5d8173c1962475071133316b},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.01864. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

How can we understand what a music foundation model has learned \textit{internally}? Most interpretability approaches, such as probing and Sparse Autoencoders (SAEs), focus on identifying individual features with minimal structural assumptions. We argue that many concepts are better understood as \textit{structured relations} rather than isolated features. This is especially prominent in music, where tonal structures are organized in the space of pitch and time. For example, concepts such as chords or keys are naturally expressed as structured sets (e.g., the 12 transpositions of a chord or the diatonic system within a key), rather than isolated features. In this study, \textbf{we shift from feature identification to structure-based analysis}, asking whether the learned inner representations of music foundation model emerge as organized structures over features. To this end, we introduce a framework that uses pitch transposition as an inductive bias to induce ordered orbits via multi-view SAE alignment. Concretely, we generate pitch-shifted input pairs and align their SAE representations to discover structured groups of pitch-related features. Experimental results show that this approach recovers orbit structures corresponding to chords, keys, and melodic patterns across two state-of-the-art music foundation models, while requiring only minimal grounding (e.g., a few anchor examples) to interpret entire concept families.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-10-02 in cs.AI).
Categorie arXiv: Sound (cs.SD) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2610.01864 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-06T10:21:28.218Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-10-02?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).