EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

MetaPCR-LLM: Universal Proof-Carrying Reasoning Across Heterogeneous Logics for Large Language Model Validation

Boris Galitsky, Vladimir Solodkin, Aleksandr Beznosikov

AI · 10.11.2026 · 7 (11) · 474

DOI: 10.3390/ai7110474 ↗

Kurz gefasst

Fachartikel in der Zeitschrift AI, online erschienen am 10.11.2026. Verfasst von Boris Galitsky und 2 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Government of Russia.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Large language models can combine inference steps governed by different logical semantics, so locally valid steps may still form an invalid reasoning chain. MetaPCR-LLM represents such chains as typed, labeled proof graphs and checks transfers between native validators. Across four constructed or transformed hallucination datasets, it improves pooled hallucination F1 from 0.77 to 0.83 relative to ValidLLP-style validation, while also improving step-validity and chain-certification decisions. On independently annotated ProcessBench, RAGTruth, and ProofWriter samples, it improves error detection, contextual hallucination detection, and query accuracy, respectively, over the matched ValidLLP-style comparator. In mixed-logic cases, disabling bridge checks reduces performance, indicating that locally accepted steps require explicit composition checks. Error analysis identifies decomposition, routing, and formalization as continuing sources of failure; specialist methods also retain advantages on some tasks. Validation incurs a modest latency increase. These findings support cross-logic composition checks in the evaluated settings, while formal guarantees remain conditional on encoded premises, supported fragments, and trusted checking procedures.

Förderung

  • Government of Russia

Zitieren

APA

Galitsky, B., Solodkin, V., & Beznosikov, A. (2026). MetaPCR-LLM: Universal Proof-Carrying Reasoning Across Heterogeneous Logics for Large Language Model Validation. AI, 7(11), 474. https://doi.org/10.3390/ai7110474

Vancouver

Galitsky B, Solodkin V, Beznosikov A. MetaPCR-LLM: Universal Proof-Carrying Reasoning Across Heterogeneous Logics for Large Language Model Validation. AI. 2026;7(11):474. doi:10.3390/ai7110474

BibTeX
@article{Galitsky2026,
  author = {Galitsky, Boris and Solodkin, Vladimir and Beznosikov, Aleksandr},
  title = {MetaPCR-LLM: Universal Proof-Carrying Reasoning Across Heterogeneous Logics for Large Language Model Validation},
  journal = {AI},
  year = {2026},
  volume = {7},
  number = {11},
  pages = {474},
  doi = {10.3390/ai7110474}
}
Kostenlos starten →