Metraipolis › archive › dc_4765cb02e024

Differentiable Fuzzy Inference Layer: A Monotone, Compositional Ordinal Reasoning Head for Large Language Models — arXiv:2609.26113

Source: arXiv (en)
Original: https://arxiv.org/abs/2609.26113
License (as archived): CC0-1.0
Archived: 2026-10-08T14:36:37.514Z
SHA-256 of the archived text: 61bd39be86af8358beccb83e1885e9cc88e6384c6a442b7e94c2ee5cd40a61d9
Stable URL of this page: https://metraipolis.com/documento/dc_4765cb02e024

How to cite

Differentiable Fuzzy Inference Layer: A Monotone, Compositional Ordinal Reasoning Head for Large Language Models — arXiv:2609.26113 — arXiv (en), https://arxiv.org/abs/2609.26113 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_4765cb02e024 (Metraipolis archive dc_4765cb02e024, sha256 61bd39be86af8358beccb83e1885e9cc88e6384c6a442b7e94c2ee5cd40a61d9)

BibTeX

@misc{dc4765cb02e024,
  title = {Differentiable Fuzzy Inference Layer: A Monotone, Compositional Ordinal Reasoning Head for Large Language Models — arXiv:2609.26113},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.26113},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_4765cb02e024; sha256 61bd39be86af8358beccb83e1885e9cc88e6384c6a442b7e94c2ee5cd40a61d9},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.26113. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

A state-of-the-art language model asked to interpret "most of most students passed" typically answers "most," though composing two instances of "most" yields a proportion closer to "some." We trace this failure to an architectural choice rather than a data deficit: standard classifier heads treat ordinal categories as independent labels, with no mechanism to respect their natural ordering or compose them algebraically. We introduce the Differentiable Fuzzy Inference Layer (DFIL), a dual-path prediction head pairing a standard classifier with a scalar-bottlenecked branch grounded in a bank of ordered membership functions. DFIL supplies two structural primitives that a label-only head cannot inherit: monotonicity in the underlying quantity, and compositional reasoning via t-norm operations without any compositional training data. The scalar branch additionally provides an interpretable interface for analyzing residual errors. We instantiate DFIL on ordinal natural-language tasks across diverse LLM families.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-23 in cs.CL).
Categorie arXiv: Computation and Language (cs.CL).
Pagina del lavoro: https://arxiv.org/abs/2609.26113 — 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 2026-10-08T13:46:15.721Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-09-23?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).