Metraipolis › archive › dc_6194c3cb2ccd
A Declarative-Procedural Perspective on Expert Routing in Bilingual Mixture-of-Experts Language Models — arXiv:2608.15102
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2608.15102
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-18T18:58:40.390Z
- SHA-256 of the archived text
8747d5e797b77ec9487760c4cede76b69ee35770caed2bf17429db7de551a7ae- Stable URL of this page
https://metraipolis.com/documento/dc_6194c3cb2ccd
How to cite
A Declarative-Procedural Perspective on Expert Routing in Bilingual Mixture-of-Experts Language Models — arXiv:2608.15102 — arXiv (en), https://arxiv.org/abs/2608.15102 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_6194c3cb2ccd (Metraipolis archive dc_6194c3cb2ccd, sha256 8747d5e797b77ec9487760c4cede76b69ee35770caed2bf17429db7de551a7ae)
BibTeX
@misc{dc6194c3cb2ccd,
title = {A Declarative-Procedural Perspective on Expert Routing in Bilingual Mixture-of-Experts Language Models — arXiv:2608.15102},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2608.15102},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_6194c3cb2ccd; sha256 8747d5e797b77ec9487760c4cede76b69ee35770caed2bf17429db7de551a7ae},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.15102. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
We investigate whether Mixture-of-Experts (MoE) language models develop linguistically structured expert routing during bilingual language acquisition. Inspired by the Declarative-Procedural framework, we analyze lexical, grammatical, and syntactic processing in a decoder-only English-German MoE Transformer trained under sequential language exposure. We construct a probe-based validation set and extract token-level routing distributions to quantify category-dependent specialisation using mutual information, routing entropy, and Jensen-Shannon distance. The curriculum-trained model exhibits a peak mutual information of 0.1148 at layer 5, indicating category-dependent differences in routing distributions across linguistic categories. Surprisingly, a no-curriculum baseline trained on mixed English-German data shows stronger aggregate specialisation, reaching a peak mutual information of 0.2599 at the same layer. These results suggest that interpretable linguistic organization emerges within MoE routing patterns even without sequential language exposure. A replication at a second training seed shows that the no-curriculum condition's specialisation concentrates on a single language whose identity is seed-dependent, whereas the curriculum consistently yields a stable, language-balanced routing profile; rather than uniformly increasing specialisation, staged bilingual exposure reduces single-language dominance. The official Github repository: this https URL Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-18 in cs.CL). Categorie arXiv: Computation and Language (cs.CL). Pagina del lavoro: https://arxiv.org/abs/2608.15102 — 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 2026-10-18T18:06:57.845Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-08-18?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).