The paper introduces FaithAlign, a theory‑scale benchmark for evaluating faithfulness in autoformalization, consisting of 460 items with 4,982 annotated phrase‑level links and a companion set FaithAlign‑Gap with 146 items lacking informal context. It frames faithfulness evaluation as a context‑aware align‑then‑label task, requiring identification of phrase‑level correspondences between informal and formal statements under specified contextual interpretations. Experiments on eight models show that GPT‑6 Astra achieves the best defect F1 of 47.8% but still struggles to recover context‑dependent correspondences, highlighting the challenge of fine‑grained, context‑aware faithfulness assessment.
arXiv:2608. 10916v1 Announce Type: cross Abstract: Autoformalisation (AF) systems map natural language reasoning steps into formal statements in a proof assistant such as Lean.
arXiv:2610.17358v1 Announce Type: cross Abstract: Formal verification proves that a program satisfies its specification, but says nothing about whether the specification captures what the author mean...
The paper introduces a reference‑free proxy‑judge framework for autoformalization, separating correctness into grounding, statement faithfulness, coverage, and inference. It presents three theoretical results linking proxy error rates to pass‑but‑wrong rates, the impact of false rejections on refinement, and a decomposition of final pass rates. The framework is instantiated with eight proxies and evaluated on ProofNet, miniF2F, e‑SNLI, and ProntoQA, showing that the predictions hold and identifying failure routes.
The paper introduces an epistemically and formally grounded ensemble (EFG) of large language model judges to evaluate autoformalization tasks in formal mathematics. It defines four criteria—logical preservation, mathematical consistency, formal quality, and formal validity—to provide a transparent, multi‑granular assessment. Experiments show that this ensemble outperforms coarse‑grained models, offering a scalable and interpretable proxy for evaluating formal mathematical reasoning.
The paper introduces GUARD, a neuro‑symbolic system that autoformalizes argumentative material by completing missing premises (guards) before formal verification. It uses large language models to generate candidate guards, Isabelle/HOL to verify them, and a contrastive test to ensure the proof depends on the original premises and does not over‑generalize. Experiments on Debatepedia and ARCT show that GUARD improves verified‑faithful scores by over 30 points and reduces leakage by about 20 points compared to prior LLM‑driven theorem proving methods.
arXiv:2606. 16541v1 Announce Type: new Abstract: Autoformalization, translating natural-language mathematics into formal proof assistants, is bottlenecked not by translation fluency but by \emph{faithfulness}: a formal statement can typecheck and be provable, yet still encode a different theorem than the source intended.
The paper investigates whether the tokens highlighted by large language models (LLMs) as key to their predictions are behaviorally consistent when those tokens are manipulated. Using sufficiency, comprehensiveness, and counterfactuality metrics, the authors evaluate closed‑source GPT‑4o‑mini and open‑source models Gemma3, Granite8B, and DeepSeek on IMDB and Steam review datasets. Results show that GPT‑4o‑mini consistently follows expected patterns, Gemma3 and Granite8B maintain sufficiency but falter under stricter interventions, and DeepSeek variants exhibit structural inconsistencies, indicating that explanation reliability varies across model families and sizes.
arXiv:2609.26035v1 Announce Type: new Abstract: Conversational agents often express answers in a uniformly confident register. We test whether expressed uncertainty, provenance-aware assertion, and e...
arXiv:2609.39228v1 Announce Type: new Abstract: We present FYAN, a human--AI harness for document-level mathematical formalization. Rather than treating theorems in isolation, FYAN coordinates an end...
SemantEq introduces a principled method for verifying that formal statements faithfully translate natural‑language mathematics by requiring each statement to transform into the other through a chain of allowed logical steps, rather than relying on trivial logical equivalence. The system evaluates proofs using Lean’s proof terms, achieving 100 % precision and 65.6 % recall on the ProofNetVerif test split, outperforming prior methods. A complementary audit tool, SemantEq‑ProofAudit, further improves recall for equivalent pairs while rejecting all non‑equivalent pairs.
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:2610.32691v1 Announce Type: new Abstract: Natural Language Autoencoders (NLAs) express a model's internal activations in the form of readable text. The verbalizer turns an activation into an En...