ACL Findings2026
LSEG: A Fine-tuning Free Method for NL2FOL via Logic-Structure and Entropy Guided Inference Controlling
Yuwei Huang, Shi Wang, Kangli Zi, Tianyu Luo, Jiang Zhixiao, Ruiheng Wang, Yufei Wang
Abstract
Large language models have shown strong generative and reasoning capabilities, yet they still struggle with natural language to first order logic (NL2FOL) translation due to logical hallucination. We propose LSEG (Logic Structure and Entropy Guided), a fine-tuning free framework designed to improve logical consistency during inference. The core idea of LSEG is to correct hidden state deviation by leveraging logical stability across logic preserving perturbations of the input. Such deviation is especially harmful in NL2FOL, as even small drifts can flip quantifier scope or logical operators, producing formulas that are syntactically valid yet logically incorrect. First, LSEG constructs perturbation-averaged direction vectors that approximate a stable logical center. Second, it derives layer-wise correction directions by contrasting original and perturbed representations. Lastly, LSEG uses an entropy-guided adaptive mechanism to inject these directions only when the model exhibits unstable or overconfident reasoning states, thereby preserving fluency while correcting logical drift. Experiments on the FOLIO and MALLS benchmarks show that LSEG consistently improves logical equivalence scores over strong baselines, despite requiring no training or parameter updates. Further evaluation on LogicLLaMA demonstrates LSEG's architecture-agnostic effectiveness.
Ask Lune
Know this paper in minutes, not an afternoon.
Lune can read every page of it and the top-tier papers that cite it, and gives you its claims, its evidence and its limits, each one cited to the paper it comes from. No skimming, no AI slop.
Lineage
Builds on10
- Language Models are Few-Shot LearnersTom B. Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah et al.NeurIPS 2020 · 64,255 citations
- Locating and Editing Factual Associations in GPTKevin Meng, David Bau, Alex Andonian, Yonatan BelinkovNeurIPS 2022 · 3,415 citations
- Discovering Language Model Behaviors with Model-Written EvaluationsEthan Perez, Sam Ringer, Kamile Lukosiute, Karina Nguyen et al.ACL Findings 2023 · 195 citations
- Logic-LM: Empowering Large Language Models with Symbolic Solvers for Faithful Logical ReasoningLiangming Pan, Alon Albalak, Xinyi Wang, William Yang WangEMNLP Findings 2023 · 145 citations
- Editing models with task arithmeticGabriel Ilharco, Marco Túlio Ribeiro, Mitchell Wortsman, Ludwig Schmidt et al.ICLR 2023 · 31 citations
Related papers
- From Hypothesis to Premises: LLM-based Backward Logical Reasoning with Selective Symbolic TranslationQingchuan Li, Mingyue Cheng, Zirui Liu, Daoyu Wang et al.AAAI 2026 · 2 citations
- Are LLMs Stable Formal Logic Translators in Logical Reasoning Across Linguistically Diversified Texts?Qingchuan Li, Jiatong Li, Zirui Liu, Mingyue Cheng et al.WWW 2026 · 1 citation
- Harnessing the Power of Large Language Models for Natural Language to First-Order Logic TranslationYuan Yang, Siheng Xiong, Ali Payani, Ehsan Shareghi et al.ACL 2024
- Do LLMs Really Struggle at NL-FOL Translation? Revealing Their Strengths via a Novel Benchmarking StrategyAndrea Brunello, Luca Geatti, Michele Mignani, Angelo Montanari et al.AAAI 2026
- Dissecting Failure Dynamics in Large Language Model ReasoningWei Zhu, Jian Zhang, Lixing Yu, Kun Yue et al.ACL 2026 · 2 citations
