Metraipolis › archive › dc_a7fa8b5d370d
Probing the phon* representations of the Mimi codebooks — arXiv:2610.35432
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.35432
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-27T08:12:23.582Z
- SHA-256 of the archived text
fc146c219fe621f42729ab584ffa5c4b8fb30e2e2585585920fd41e861c80d7e- Stable URL of this page
https://metraipolis.com/documento/dc_a7fa8b5d370d
How to cite
Probing the phon* representations of the Mimi codebooks — arXiv:2610.35432 — arXiv (en), https://arxiv.org/abs/2610.35432 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_a7fa8b5d370d (Metraipolis archive dc_a7fa8b5d370d, sha256 fc146c219fe621f42729ab584ffa5c4b8fb30e2e2585585920fd41e861c80d7e)
BibTeX
@misc{dca7fa8b5d370d,
title = {Probing the phon* representations of the Mimi codebooks — arXiv:2610.35432},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.35432},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_a7fa8b5d370d; sha256 fc146c219fe621f42729ab584ffa5c4b8fb30e2e2585585920fd41e861c80d7e},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.35432. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Neural audio codecs are increasingly used to encode speech for integration with large language models. This paper investigates the phonetic representations in the first codebook (CB0) of the Mimi codec. ABX style evaluation is poorly suited to Mimi fixed 80 ms tokenization window, which often spans multiple phonemes; we instead use forced alignment, entropy analysis, and Phone-Normalized Mutual Information (PNMI). Using TIMIT speech (58/55 speakers, SA1/SA2), individual tokens are strongly associated with specific phonemes (PNMI = 0.84, above random), yet phonemes are realized by multiple competing tokens rather than a one-to-one mapping. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.CL, nuovi lavori del giorno). Categorie arXiv: Computation and Language (cs.CL). Pagina del lavoro: https://arxiv.org/abs/2610.35432 — 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 2026-10-27T07:56:00.266Z dalla pagina https://arxiv.org/list/cs.CL/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).