Metraipolis › archive › dc_48b7aba655e4

Data Scarcity and Model Sparsity: Mixtures-of-Experts Overfit More to Repeated Data — arXiv:2609.11917

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.11917
License (as archived)
CC0-1.0
Archived
2026-10-16T05:44:34.250Z
SHA-256 of the archived text
94dda7f66a7b1569a8a0b4f081ce07e5d85e9b2830d83064ab21a83c0f1e39c0
Stable URL of this page
https://metraipolis.com/documento/dc_48b7aba655e4

How to cite

Data Scarcity and Model Sparsity: Mixtures-of-Experts Overfit More to Repeated Data — arXiv:2609.11917 — arXiv (en), https://arxiv.org/abs/2609.11917 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_48b7aba655e4 (Metraipolis archive dc_48b7aba655e4, sha256 94dda7f66a7b1569a8a0b4f081ce07e5d85e9b2830d83064ab21a83c0f1e39c0)
BibTeX
@misc{dc48b7aba655e4,
  title = {Data Scarcity and Model Sparsity: Mixtures-of-Experts Overfit More to Repeated Data — arXiv:2609.11917},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.11917},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_48b7aba655e4; sha256 94dda7f66a7b1569a8a0b4f081ce07e5d85e9b2830d83064ab21a83c0f1e39c0},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.11917. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

As the supply of human-written text is exhausted, it has become standard practice to repeat language model training data. Prior work has studied data repetition for densely activated Transformers, but the effects of data repetition remains largely unexplored for recently dominant sparse architectures such as Mixture-of-Experts (MoE), despite their increased compute efficiency. We vary data repetition rates across single- and multi-domain data mixes, and across MoE settings, including expert count and granularity. We consistently find, for models ranging from 80M to 1B active (8.5B total) parameters, that MoEs degrade more rapidly under data repetition. This effect increases with sparsity, dictated by total rather than active parameters. While 80M dense models can repeat data over 8x with minimal degradation, MoEs instead begin to suffer at 4x, and deteriorate rapidly, ceding their performance benefits in all-unique data settings to underperform dense models after 32x. We experiment with existing regularization methods as a potential remedy. We find that some methods, such as dropout, can mitigate overfitting. In particular, with strong masking-based regularization, MoEs are able to outperform dense models even when data is repeated more than 64 times. However, no method fully matches the performance of all-unique training data. Finally, we analyze internal mechanisms correlated with MoE overfitting in high repetition regimes, and find that MoE routing universally stabilizes early in training, and that expert specialization correlates with overfitting to repeated data. In sum, our work addresses the adverse interactions between sparsity and data repetition: we present evidence for the core mechanisms of overfitting and its potential remediation, and suggest promising avenues for future methods to reduce over-specialization in model parameters by disrupting memorization patterns.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-11 in cs.CL).
Categorie arXiv: Machine Learning (cs.LG) ; Computation and Language (cs.CL).
Pagina del lavoro: https://arxiv.org/abs/2609.11917 — 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 2026-10-16T05:23:57.052Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-09-11?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).