arXiv:2607. 17047v1 Announce Type: cross Abstract: LLM constraint reasoners are often evaluated near the random-SAT phase transition, confounding density and solver hardness.
arXiv:2607. 29431v1 Announce Type: new Abstract: Large language models increasingly generate optimization models from natural language, but existing evaluation often reduces a generated model and its ground truth to a single equivalent/not-equivalent verdict or an execution-success rate--labels that are neither independently checkable nor faithful to the multiple distinct senses in which two formulations can agree.
arXiv:2608. 14771v1 Announce Type: new Abstract: Making language models solve constraint problems reliably often means having them translate the problem into a formal specification and delegating the search to a sound solver.
arXiv:2608. 13228v1 Announce Type: new Abstract: Agent harnesses combine retrieval, routing, state, provenance, and verification, but locally successful components may disagree on shared state.
arXiv:2608. 04552v1 Announce Type: cross Abstract: Black-box language-model reliability is commonly pursued by sampling, prompting, voting, verifying, or iteratively revising individual answers.
arXiv:2608. 10420v1 Announce Type: new Abstract: Reasoning shortcuts are solutions of a neurosymbolic system's rules that produce correct predictions through unintended concepts.
arXiv:2608. 12599v1 Announce Type: new Abstract: Multi-turn dialogues let users revoke constraints as easily as impose them, but revocation does not reliably take effect: models keep enacting withdrawn requirements (occasionally beneath comments asserting their removal), a failure we call \emph{behavioral relapse}, or revocation inertia.
arXiv:2608. 08008v1 Announce Type: new Abstract: Process reward models (PRMs) score intermediate reasoning steps and are widely used for search, ranking, and training, but optimization can exploit these learned proxies by increasing reward while turning correct reasoning into incorrect reasoning.
SemVerBench is a benchmark that evaluates how well large language models (LLMs) understand and apply version-constraint resolution semantics, such as determining whether a version satisfies constraints like ^1.2.3 or >=2.0. The study finds that many models struggle with certain corner cases, with GPT‑5.1 performing poorly while Claude and Opus perform much better. The authors suggest that the failures stem from an activation/application gap rather than a lack of knowledge, and recommend that coding agents delegate version resolution to a dedicated resolver tool.
arXiv:2608. 12321v1 Announce Type: cross Abstract: When a salient surface cue competes with an implicit feasibility constraint, LLMs often fail -- but aggregate accuracy conflates genuine constraint inference with conservative defaulting.
The paper introduces Constraint‑Guided Enterprise Data Mapping (CGM), a neuro‑symbolic approach that uses schema‑grounded admissibility constraints to steer large language models (LLMs) in aligning enterprise data. CGM operates in three stages: defining constraints with metadata, generating candidates under relaxed constraints to ensure feasibility, and ranking them with a bounded LLM. Experiments show that hard constraints dramatically reduce candidate space and improve F1 scores, enabling small models to match or surpass large LLMs at a fraction of the cost while reducing expert effort.
The paper establishes a theoretical bound on the cost of enforcing a physics prior in machine learning models, showing that the excess risk of a shape‑constrained hypothesis class is always bounded by the excess risk of an ablated model that ignores the prior. Empirical tests on an ordinal wildfire‑severity task confirm that a constrained model can never be outperformed by its own ablation, and that the cost of the prior is protocol‑dependent and can be quantified using a self‑calibrating floor. The authors also propose a two‑fit screening method to reject unidentifiable experiments before training a constrained model.