Metraipolis › archive › dc_b37ccae304fe

Probing for Long-Horizon Deductive Reasoning Capabilities in Language Models with Prolog — arXiv:2610.11592

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.11592
License (as archived)
CC0-1.0
Archived
2026-10-09T06:13:15.582Z
SHA-256 of the archived text
038659bab80ca15256bc073ae3e21ad4ba8fbcea19887acf7061eac53c75f9ea
Stable URL of this page
https://metraipolis.com/documento/dc_b37ccae304fe

How to cite

Probing for Long-Horizon Deductive Reasoning Capabilities in Language Models with Prolog — arXiv:2610.11592 — arXiv (en), https://arxiv.org/abs/2610.11592 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_b37ccae304fe (Metraipolis archive dc_b37ccae304fe, sha256 038659bab80ca15256bc073ae3e21ad4ba8fbcea19887acf7061eac53c75f9ea)
BibTeX
@misc{dcb37ccae304fe,
  title = {Probing for Long-Horizon Deductive Reasoning Capabilities in Language Models with Prolog — arXiv:2610.11592},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.11592},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_b37ccae304fe; sha256 038659bab80ca15256bc073ae3e21ad4ba8fbcea19887acf7061eac53c75f9ea},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.11592. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Current frontier LLMs can theoretically process long contexts with 1M tokens or more. But to what extent can they go beyond simple retrieval and perform deeper reasoning over such long contexts? We empirically investigate long-horizon reasoning capabilities of LLMs, focusing on deductive logic expressed in Prolog. We construct ProloNg, a synthetic testbed to probe Prolog Long Reasoning, which systematically varies the complexity (reasoning depth) of problems, where the hardest case has a reasoning depth of 22 and 62k context length. We study 8 reasoning models across 5 families of frontier LLMs, and find that performance degrades substantially as reasoning depth grows, with the majority of models approaching chance beyond depth 10.

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/2610.11592 — 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 2026-10-09T05:47:15.223Z dalla pagina https://arxiv.org/list/cs.CL/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).