Metraipolis › archive › dc_fcb8035aaf22

SCOPE: Compiling Natural-Language Policies into Neurosymbolic Predicate Programs — arXiv:2610.31062

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.31062
License (as archived)
CC0-1.0
Archived
2026-10-25T08:32:50.798Z
SHA-256 of the archived text
5267a58dc295520c071672fa4c71c959674dbb75c79d97e3cd6b263b652fd430
Stable URL of this page
https://metraipolis.com/documento/dc_fcb8035aaf22

How to cite

SCOPE: Compiling Natural-Language Policies into Neurosymbolic Predicate Programs — arXiv:2610.31062 — arXiv (en), https://arxiv.org/abs/2610.31062 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_fcb8035aaf22 (Metraipolis archive dc_fcb8035aaf22, sha256 5267a58dc295520c071672fa4c71c959674dbb75c79d97e3cd6b263b652fd430)
BibTeX
@misc{dcfcb8035aaf22,
  title = {SCOPE: Compiling Natural-Language Policies into Neurosymbolic Predicate Programs — arXiv:2610.31062},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.31062},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_fcb8035aaf22; sha256 5267a58dc295520c071672fa4c71c959674dbb75c79d97e3cd6b263b652fd430},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.31062. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Policy enforcement for LLM agents has shifted from single-turn content moderation to compound, application-specific policies evaluated over multi-turn transcripts and tool-call trajectories. Fine-tuned guardrail models are efficient but degrade on compound policies, and while LLM-as-a-judge approaches achieve higher accuracy, they incur high inference cost and tend to produce unstable predictions. Neurosymbolic methods offer a natural alternative, decomposing the policy into narrow neural decisions composed under explicit logic. However, existing symbolic layers lack the required expressiveness and are optimized against fixed data rather than the policy itself. To address this, we introduce SCOPE, which compiles a policy into an executable program of neural predicates executed by small models, deterministic extractors, and code that composes their outputs. The program is optimized jointly with its training data, using adversarial generation of counterexamples that target where the program deviates from the policy. We evaluate SCOPE on a benchmark of real-world policy tasks against prompt optimization, task decomposition, and workflow generation methods, achieving the highest average accuracy across tasks and models. Programs optimized with the weakest executor transfer to other models without further optimization. This allows them to be deployed in production on inexpensive models, where they reach accuracy comparable to frontier reasoning models. Unlike these monolithic models, SCOPE produces substantially more stable predictions, and each verdict is interpretable through the predicate outputs that produced it. Additionally, we validate the key components of our method through an ablation study. SCOPE thus enables reliable, interpretable, and efficient policy enforcement at production scale.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Artificial Intelligence (cs.AI) ; Computation and Language (cs.CL).
Pagina del lavoro: https://arxiv.org/abs/2610.31062 — 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 2026-10-25T08:31:13.060Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).