Pith. sign in

REVIEW 5 major objections 5 minor 57 references

Semiotic logical hexagon theory for LLM logical reasoning

T0 review · 5 major / 5 minor · reviewed 2026-08-15 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read This paper argues that laying out a statement's six logically opposing meanings before deduction raises LLM reasoning accuracy by about 2.4–2.7 points across five benchmarks.

desk verdict A sound formal hexagon paper whose empirical bridge is unproven: the headline accuracy deltas rest on single unseeded runs, a self-authored baseline, and an ablation that exactly reproduces the baseline. read the letter →

arxiv 2607.21933 v1 pith:VVS7CMXS submitted 2026-07-24 cs.AI

Yunyao Zhang , Xinglang Zhang , Zeliang Chen , Junqing Yu , Zikai Song This is my paper · ORCID
classification cs.AI
keywords largelanguagemodelslogicalreasoninghexagonsemanticorganizationreflectiveverificationfirst-orderlogicneuro-symboliccomplexity
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper argues that large language models reason better when the hidden meaning of a target statement is laid out as a complete logical hexagon before any deduction begins. It claims that a four-position semantic square leaves out two load-bearing meanings—the uniform alternative (all objects are W or all are not W) and the mixed alternative (some are W and some are not)—and that adding them improves reasoning reliability across five benchmarks and three model backbones. If the claim holds, the main cause of LLM logical failure is not the deduction step but the incomplete semantic organization that precedes it, and a six-position structure of opposing meanings delays the accuracy collapse that appears as reasoning complexity grows.

What carries the argument

The load-bearing object is the logical hexagon instantiated in first-order logic: starting from a subject predicate M(x) and a property W(x), the six positions are A = ∀x(M(x)→W(x)), E = ∀x(M(x)→¬W(x)), I = ∃x(M(x)∧W(x)), O = ∃x(M(x)∧¬W(x)), U = A∨E (uniform class), and Y = I∧O (mixed class). Its effect comes from the four relation families of Proposition 2: A, E, Y are pairwise contraries; U, I, O are pairwise subcontraries; A contradicts O, E contradicts I, U contradicts Y; and six subalternations hold. These relations, registered as constraints before reasoning, do the work in the reflective verification stage: a True verdict at one contrary position forces False at another, a False verdict at one subcontrary position forces True at another, and a True superordinate forces True at its subordinate. The ablation study identifies the contradictory anchor ¬ρ(Q) and the contrary/subcontrary checks as the most valuable pieces, with the composite positions U and Y contributing less.

What would settle it

Take a set of reasoning problems whose target statements involve empty subject classes (e.g., 'All unicorns are white') or inherently relational content that cannot be expressed as one subject–property pair, run HexLogicAgent on them, and check whether the final True/False/Unknown verdicts violate the hexagon's predicted relations—for an empty class, both A and E are vacuously true, which would break the pairwise-contrariety rule the verifier relies on.

Watch

Extended reading notes

Core claim

The central claim is Proposition 2: under existential import, the six propositions A (all M are W), E (all M are not W), I (some M are W), O (some M are not W), U = A∨E, and Y = I∧O form a complete logical hexagon whose contradiction, contrariety, subcontrariety, and subalternation relations hold by construction. The paper instantiates this in HexLogicAgent, which maps a natural-language target to a normalized tuple ⟨x, M, W⟩, constructs all six positions, proves the anchor position and its contradiction, and then uses the hexagon relations to resolve conflicting verdicts into a final True/False/Unknown answer. Its experiments report consistent average accuracy gains of +2.41 to +2.74 points over the strongest baseline across three frontier backbones, with a 100% success rate on a semantic-understanding case set where the four-position predecessor scored 13%. The paper's second observation is that most remaining errors are uncertainty-calibration failures (True→Unknown, False→Unknown) rather than direct polarity reversals, which it reads as evidence that incomplete semantic representation, not deductive inference, is the dominant failure source.

Load-bearing premise

The framework stands on the assumption that every target statement can be faithfully squeezed into a single subject–property pair ⟨x, M, W⟩ and that the subject class is non-empty; if a real statement resists that normalization or talks about an empty class, the hexagon's contrariety and subcontrariety relations no longer force correct answers.

Editorial extensions

If this is right

  • On the five tested benchmarks, switching from the semiotic square to the full hexagon adds +2.41 to +2.74 average accuracy points on three frontier backbones, with the largest gains on ProofWriter and ProverQA.
  • Explicitly modeling complete semantic opposition delays the logical phase transition: the accuracy gap between HexLogicAgent and prompting baselines widens as logical complexity (LoCM) increases, rather than shrinking.
  • If incomplete semantic organization is the dominant failure source, then adding more semantic structure before deduction should continue to pay off on harder three-way reasoning tasks, not just on synthetic proof chains.
  • The framework's runtime stays on the Pareto frontier of accuracy versus per-sample time, so the accuracy improvement is not bought with dramatically more computation than prior symbolic methods.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • Editorial extension: the same hexagon structure could be tested as a general pre-processing layer for contradiction detection or fact-checking, where deciding whether a claim is opposite, contrary, or merely a local counterexample is the core difficulty.
  • Editorial extension: because the paper's gains rely on a single normalized tuple, a natural test is whether multi-predicate relational reasoning benefits from composing hexagons over each predicate or from a richer tensor of semantic positions.
  • Editorial extension: the ablation shows the composite U and Y positions contribute least; a focused comparison against a 'triangle' constraint system (only contradictions plus one contrary pair) would show whether the full hexagon is necessary or just sufficient.
  • Editorial extension: the paper compares against fixed baselines, not against an ablation that uses the same reflective-verification prompts with hexagon labels removed; such a control would isolate whether the semantic structure itself, rather than additional verification tokens, drives the gain.
Share X Bluesky LinkedIn Reddit HN

Signed reviews

No signed human review yet.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

5 major / 5 minor

Summary. The paper proposes HexLogicAgent, a reasoning framework that organizes a natural-language target proposition into the six positions {A, E, I, O, U, Y} of the logical hexagon before performing LLM-based deduction, and then verifies the resulting verdicts against the hexagon's contradiction, contrariety, subcontrariety, and subalternation constraints. The formal core is Proposition 2, which states that, under classical FOL with existential import over the subject class, the six positions form a complete hexagonal closure with the four relation types. Empirically, the paper reports average accuracy gains of +2.66, +2.74, and +2.41 points over its own LogicAgent baseline across three backbone LLMs and five benchmarks (FOLIO, ProntoQA, ProofWriter, ProverQA, RepublicQA). It also claims that incomplete semantic representations are a major source of LLM reasoning failures, that the hexagon delays performance degradation under increasing logical complexity, and that HexLogicAgent improves semantic understanding from 13% to 100% on representative cases.

Significance. If the empirical claims hold, the paper makes a useful contribution by importing an established logical structure, the Blanché hexagon, into LLM-based reasoning and by proposing a clean three-stage pipeline with a sound formal basis. The formal propositions are correct under their stated assumptions, and the framework is well-motivated by the known incompleteness of the semiotic square for representing existential and mixed-state meanings. The paper is less novel in its formal apparatus, which is independently known in the logic literature, than in its application. The main strengths are the explicit formalization, the clear comparison against a relevant prior method, and the ablation of hexagon components. However, the empirical support for the central claim is currently not statistically grounded: the headline gains are small, come from single unseeded runs, and are measured against the authors' own baseline and benchmark, with no released prompts or code. The paper also contains an internal inconsistency between the abstract's claim about incomplete semantic representations as the major failure source and the error analysis that attributes errors to uncertainty calibration.

major comments (5)
  1. [§4.2, Table 1] The central empirical claim, that HexLogicAgent 'consistently improves reasoning reliability' with average gains of +2.66, +2.74, and +2.41 points over LogicAgent, rests on single runs with no error bars, random seeds, sampling temperatures, or significance tests. Table 3 shows that removing U and Y exactly reproduces the LogicAgent row (86.77 on DeepSeek-V3.2 and 83.55 on Qwen3-30B-A3B), so the entire headline delta over the strongest baseline is a 2-3 point average that may be within run-to-run noise. Please report multiple seeds, variance, and a statistical comparison (e.g., paired tests or bootstrap across benchmarks), and release the prompts and code needed to reproduce the results.
  2. [Abstract vs. §4.4.2] The abstract claims that 'incomplete semantic representations, rather than deductive inference itself, are a major source of logical reasoning failures in LLMs,' but the paper's own error analysis does not support this. Section 4.4.2 finds that 64.4% of errors are 'reasoning insufficiency' (True/False predicted as Unknown), and item (3) concludes that 'uncertainty calibration [is] the main remaining source of error in HexLogicAgent.' No evidence connects errors to incomplete semantic representations. Either test this claim directly with a representation-completeness analysis or revise the claim to match the evidence.
  3. [§4.4.1, Figure 6] The phase-transition analysis is not reproducible because the LoCM (logical complexity) metric is never defined. The paper cites the authors' prior work [44] without giving a formal definition of LoCM or explaining how complexity bins are constructed. Without this definition, the observation that HexLogicAgent 'delays logical phase transitions' cannot be evaluated or replicated. Please define LoCM precisely or remove the claim.
  4. [§3.2, Figure 4] The framework's load-bearing step is predicate identification, which normalizes the target proposition into a subject-property tuple ⟨x, M, W⟩, yet the paper provides no evaluation of how often or how accurately this extraction succeeds. The 'semantic understanding success rate' of 100% (HexLogicAgent) versus 13% (LogicAgent) and 0% (other methods) is computed on 'representative semantic understanding cases' with no description of how these cases were selected, how many there are, or how success is scored. This overstates the contribution of semantic structuring. Please specify the evaluation protocol for these cases and, ideally, report extraction accuracy on the full benchmarks.
  5. [§3.4, Tables 2-3] The reflective verification stage is a core component of the framework, but its independent accuracy contribution is not measured. Figure 8(b) lists a 'w/o Reflection' ablation for runtime, yet Tables 2 and 3 do not include an accuracy row for removing the entire reflective verification stage; the closest rows ('w/o Hexagon', 'w/o Ctr/Subctr') also remove structural components. Because reflective verification adds significant computation (44.4% runtime reduction when removed), the observed gains could come from the extra verification compute rather than from the hexagon structure itself. Please report the accuracy of a 'w/o Reflection' variant to separate these effects.
minor comments (5)
  1. [§4.1] RepublicQA is a benchmark introduced in the authors' prior LogicAgent paper [34], and LogicAgent is the authors' own strongest baseline. This makes the comparison partially self-referential; the paper should acknowledge this and note that independent evaluation on external benchmarks would strengthen the claims.
  2. [Figure 6] The figure caption and axis labels are garbled (e.g., 'LoCM' is not expanded and some characters are corrupted), making the phase-transition curves difficult to interpret. Please provide a clean figure with a defined x-axis.
  3. [§4.4.2] The failure-mode categories (reasoning insufficiency, overconfidence, logical fallacy) are not formally defined, and the classification procedure is not described. Please provide the criteria used to assign errors to these categories.
  4. [Introduction] The introduction states that the article provides 'full implementation details,' but no code or prompts are released or referenced. Please state where the implementation will be made available and include the actual prompt templates in an appendix.
  5. [§4.3, Table 3] The claim that U and Y contribute less than other components is based on a single ablation row; please include variance or multiple runs for these ablations as well, since the differences are small and may not be significant.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the hexagon formalism is proved in-paper from explicit FOL definitions, and the main empirical claim is supported by multiple external benchmarks and external baselines; the self-authored LogicAgent comparison is an incremental ablation, not a definitional input.

full rationale

The paper's formal core is self-contained: Proposition 2 defines A, E, I, O, U, Y in first-order logic and proves contradiction, contrariety, subcontrariety, and subalternation from those definitions under existential import. The logical hexagon itself is explicitly attributed to prior work by Béziau and Moretti, so the theory is not presented as a novel derivation from a self-citation. The extension from the semiotic square to the hexagon is argued by Proposition 1, which shows that I and O are not contraries and that I∧O and A∨E are not represented in the square; this is a mathematical demonstration rather than a circular redefinition. The empirical claim does not reduce to a fitted input called a prediction: no trainable parameters are fit to the benchmarks, and the reported gains are measured accuracies from a fixed prompt-based pipeline. The ablation 'w/o U,Y' exactly recovering the LogicAgent averages is expected by construction, because HexLogicAgent is defined as LogicAgent plus the two composite positions; that is a controlled comparison to the authors' own prior method, not a self-fulfilling prediction. The evaluation also includes four externally established benchmarks (FOLIO, ProntoQA, ProofWriter, ProverQA) and multiple external baselines (CoT, ToT, CR, DetermLR, LogicLM, SymbCoT, Aristotle), so the central claim does not stand or fall on the self-introduced RepublicQA or the self-authored LogicAgent baseline. The paper's secondary observation that incomplete semantic representations are the main failure source is weakened by its own error analysis, which attributes most errors to uncertainty calibration; however, that is an internal inconsistency or correctness concern, not a circularity. The absence of seeds, error bars, released prompts, and code is a reproducibility limitation rather than evidence that the derivation is circular. Overall, no load-bearing step was found that reduces to its own inputs by construction or by self-citation chain.

Assumptions & free parameters 0 free parameters · 4 assumptions · 0 invented entities

The formal framework is elementary first-order logic; its correctness rests on existential import and classical bivalence. The mapping from natural language to a single subject-property tuple is a modeling assumption not guaranteed by the formal theory. No numeric parameters are fitted to data; the main empirical decisions are prompt and pipeline choices not specified as exact constants.

assumptions (4)
  • domain assumption Existential import: the subject class is non-empty, ∃x M(x), whenever opposition relations are analyzed.
    Invoked in Section 2.1 and used in Proposition 2; without it, A and E can both be vacuously true and I and O can both be false, breaking contrariety and subcontrariety.
  • standard math Classical bivalence: every proposition is either true or false.
    Used in the proof of subcontrariety in Section 2.3 to conclude that I and O cannot both be false under existential import.
  • ad hoc to paper Target natural-language propositions can be normalized to a single subject-property tuple ⟨x, M, W⟩.
    Section 3.2, Predicate Identification; the entire hexagon construction depends on this normalization being faithful for arbitrary natural-language statements.
  • ad hoc to paper The six FOL positions {A, E, I, O, U, Y} form a complete closure of the relevant semantic space for a proposition.
    Section 2.3 and Proposition 2; completeness is claimed for this particular instantiation, not proven for all natural-language semantic relations.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Semiotic logical hexagon theory for LLM logical reasoning." pith.science (2026). https://pith.science/paper/VVS7CMXS

@misc{pith2026260721933,
  author       = {Pith},
  title        = {Pith review of: Semiotic logical hexagon theory for LLM logical reasoning},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/VVS7CMXS}},
  note         = {Machine review of arXiv:2607.21933}
}
read the original abstract

Large language models (LLMs) have become powerful tools for language understanding and logical reasoning. However, they still make mistakes when a problem requires both understanding meaning and following logic. A key reason is that natural-language statements often carry implicit semantic relations before any formal reasoning begins. If these hidden meanings are not properly organized, the model may reach incorrect conclusions even when the subsequent reasoning process appears logically valid. Existing methods improve reasoning through decomposition, symbolic translation, external solvers, or self-verification, but pay comparatively less attention to the semantic structure on which reasoning depends. In this paper, we further investigate how semantic organization influences logical reasoning in LLMs. To this end, we propose HexLogicAgent, a framework that first organizes the meaning of natural-language statements and then guides logical reasoning through structured verification. In our investigation, we also make two observations. First, incomplete semantic representations, rather than deductive inference itself, are a major source of logical reasoning failures in LLMs. Second, explicitly modeling the complete structure of semantic opposition substantially delays the degradation of reasoning performance as logical complexity increases. Experiments on challenging logical reasoning benchmarks demonstrate that HexLogicAgent consistently improves reasoning reliability across multiple LLMs. The core idea is supported by a logical hexagon theory, which explains why a complete structure of opposing meanings is necessary for reliable reasoning.

Figures

Figures reproduced from arXiv: 2607.21933 by the authors.

Figure 1
Figure 1. Comparison of representative LLM reasoning paradigms. (1) Linear Reasoning (LR) follows a single reasoning chain, (2) Aggregative Reasoning (AR) aggregates multiple paths, and (3) Symbolic Reasoning (SR) couples LLMs with symbolic modules. (4) Semiotic Reasoning (SeR), instantiated by HexLogicAgent, organizes proposition meanings through semiotic structures before deduction and verification. into three categories: (… view at source ↗
Figure 2
Figure 2. Comparison between the semiotic square and the logical hexagon. The semiotic square captures contrary, contradictory, and subaltern relations, while the logical hexagon extends it with existential and mixed-state positions. This enables HexLogicAgent to model subcontrary relations, distinguish local counterexamples from global negation, and represent partial proposition satisfaction. 2.2. Greimas’ Semiotic Square an… view at source ↗
Figure 3
Figure 3. Overview of HexLogicAgent. The agent processes a natural-language reasoning problem through three stages. (1) Hexagonal Semantic Structuring extracts the predicate tuple  (𝑄) = ⟨𝑥,𝑀, 𝑊 ⟩, constructs the logical hexagon (𝑄) = {𝐴, 𝐸, 𝐼, 𝑂, 𝑈, 𝑌 }, and maps the target proposition to its anchor position 𝜌(𝑄). (2) Logical Reasoning formalizes the premises into Γ𝑃 , plans two proof goals for 𝜌(𝑄) and ¬𝜌(𝑄), and derives … view at source ↗
Figures from the paper (6 more)
Figure 4
Figure 4. Figure 4: Semantic-understanding success rate. The figure compares the semantic-understanding success rates of the other reasoning methods, LogicAgent, and HexLogicAgent, which achieve 0%, 13%, and 100%, respectively. conventional reasoning methods fail to correctly capture the …
Figure 5
Figure 5. Figure 5: Ablation drops and dataset-wise effect of planning removal. Panels (a) and (b) show absolute average accuracy drops under mechanism-level and hexagonal-structure ablations, where larger |ΔAVG| indicates a more important component. Panel (c) shows the dataset-wise accur…
Figure 6
Figure 6. Figure 6: Phase transition curves across LoCM. Panels (a)–(c) compare Direct prompting, CoT, and LogicAgent on DeepSeek-V3.2, Qwen2.5-32B-Instruct, and Qwen3-30B-A3B, respectively. Panels (d)–(e) compare LogicAgent with representative training-based methods on Qwen2.5-7B-Instruc…
Figure 7
Figure 7. Figure 7: Failure-mode analysis of HexLogicAgent. Panel (a) shows the overall composition of erroneous predictions, categorized as reasoning insufficiency, overconfidence, and logical fallacy. Panel (b) reports the normalized failure composition for each benchmark, revealing tha…
Figure 8
Figure 8. Figure 8: Time-efficiency analysis of HexLogicAgent. Panel (a) compares the per-sample solving time of HexLogicAgent and existing reasoning methods on a logarithmic scale. Panel (b) reports the per-sample solving time under different component ablations, with percentages denotin…
Figure 9
Figure 9. Figure 9: Case study illustrating how HexLogicAgent distinguishes a local counterexample from a universal negation. Zhang et al.: Preprint submitted to Elsevier Page 17 of 19 [PITH_FULL_IMAGE:figures/full_fig_p017_9.png]

Review history (2 revisions) →

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

57 extracted references · 18 canonical work pages

  1. [44]

    Logical phase transitions: Understanding collapse in llm logical reasoning, 2026

    Xinglang Zhang, Yunyao Zhang, ZeLiang Chen, Junqing Yu, Wei Yang, and Zikai Song. Logical phase transitions: Understanding collapse in llm logical reasoning, 2026

    2026

  2. [1]

    Deepseek-r1incentivizesreasoninginllmsthroughreinforcementlearning.Nature,645(8081): 633–638, 2025

    DayaGuo,DejianYang,HaoweiZhang,etal. Deepseek-r1incentivizesreasoninginllmsthroughreinforcementlearning.Nature,645(8081): 633–638, 2025

    work page 2025

  3. [2]

    Gpt-4o system card.arXiv preprint arXiv:2410.21276, 2024

    Aaron Hurst, Adam Lerer, Adam P Goucher, Adam Perelman, Aditya Ramesh, Aidan Clark, AJ Ostrow, Akila Welihinda, Alan Hayes, Alec Radford, et al. Gpt-4o system card.arXiv preprint arXiv:2410.21276, 2024

    arXiv 2024

  4. [3]

    An Yang, Bowen Yu, Chengyuan Li, Dayiheng Liu, Fei Huang, Haoyan Huang, Jiandong Jiang, Jianhong Tu, Jianwei Zhang, Jingren Zhou, et al. Qwen2. 5-1m technical report.arXiv preprint arXiv:2501.15383, 2025

    arXiv 2025

  5. [4]

    White, Longqi Yang, Reid Andersen, Georg Buscher, Dhruv Joshi, and Nagu Rangan

    Mengting Wan, Tara Safavi, Sujay Kumar Jauhar, Yujin Kim, Scott Counts, Jennifer Neville, Siddharth Suri, Chirag Shah, Ryen W. White, Longqi Yang, Reid Andersen, Georg Buscher, Dhruv Joshi, and Nagu Rangan. Tnt-llm: Text mining at scale with large language models. InProceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, KDD ...

    arXiv 2024

  6. [5]

    Connecting the dots: Llms can infer and verbalize latent structure from disparate training data

    Johannes Treutlein, Dami Choi, Jan Betley, Sam Marks, Cem Anil, Roger Grosse, and Owain Evans. Connecting the dots: Llms can infer and verbalize latent structure from disparate training data. In A. Globerson, L. Mackey, D. Belgrave, A. Fan, U. Paquet, J. Tomczak, and C. Zhang, editors,Advances in Neural Information Processing Systems, volume 37, pages 140...

    work page doi:10.52202/079017-4465 2024

  7. [6]

    Properties and challenges of LLM-generated explanations

    Jenny Kunz and Marco Kuhlmann. Properties and challenges of LLM-generated explanations. In Su Lin Blodgett, Amanda Cercas Curry, Sunipa Dev, Michael Madaio, Ani Nenkova, Diyi Yang, and Ziang Xiao, editors,Proceedings of the Third Workshop on Bridging Human– Computer Interaction and Natural Language Processing, pages 13–27, Mexico City, Mexico, June 2024. ...

    doi:10.18653/v1/2024.hcinlp-1.2 2024

  8. [7]

    Openagi: When llm meets domainexperts

    Yingqiang Ge, Wenyue Hua, Kai Mei, jianchao ji, Juntao Tan, Shuyuan Xu, Zelong Li, and Yongfeng Zhang. Openagi: When llm meets domainexperts. InA.Oh,T.Naumann,A.Globerson,K.Saenko,M.Hardt,andS.Levine,editors,AdvancesinNeuralInformationProcessing Systems,volume36,pages5539–5568.CurranAssociates,Inc.,2023. URLhttps://proceedings.neurips.cc/paper_files/paper...

    arXiv 2023

Show all 57 references
  1. [8]

    CANDLE: Iterative conceptualization and instantiation distillation from large language models for commonsense reasoning

    Weiqi Wang, Tianqing Fang, Chunyang Li, Haochen Shi, Wenxuan Ding, Baixuan Xu, Zhaowei Wang, Jiaxin Bai, Xin Liu, Cheng Jiayang, Chunkit Chan, and Yangqiu Song. CANDLE: Iterative conceptualization and instantiation distillation from large language models for commonsense reason...

    2024

  2. [9]

    Mathcoder: Seamless code integration in llms for enhanced mathematical reasoning.arXiv preprint arXiv:2310.03731, 2023

    Ke Wang, Houxing Ren, Aojun Zhou, Zimu Lu, Sichun Luo, Weikang Shi, Renrui Zhang, Linqi Song, Mingjie Zhan, and Hongsheng Li. Mathcoder: Seamless code integration in llms for enhanced mathematical reasoning.arXiv preprint arXiv:2310.03731, 2023

    arXiv

  3. [10]

    Deepse(3)-equivariantgeometricreasoningforprecise placement tasks.arXiv preprint arXiv:2404.13478, 2024

    BenEisner,YiYang,TodorDavchev,MelVecerik,JonathanScholz,andDavidHeld. Deepse(3)-equivariantgeometricreasoningforprecise placement tasks.arXiv preprint arXiv:2404.13478, 2024

    2024 arXiv

  4. [11]

    Critical and reflective thinking: A philosophical perspective

    Richard W Paul. Critical and reflective thinking: A philosophical perspective. InDimensions of thinking and cognitive instruction, pages 445–494. Routledge, 2013

    2013

  5. [12]

    Coupling macro dynamics and micro states for long-horizon social simulation, 2026

    Yunyao Zhang, Yihao Ai, Zuocheng Ying, Qirui Mi, Junqing Yu, Wei Yang, and Zikai Song. Coupling macro dynamics and micro states for long-horizon social simulation, 2026

    2026

  6. [13]

    Intervensim: Intervention- aware social network simulation for opinion dynamics, 2026

    Yunyao Zhang, Zuocheng Ying, Xinglang Zhang, Junqing Yu, Peng Fang, Xu Chen, Wei Yang, and Zikai Song. Intervensim: Intervention- aware social network simulation for opinion dynamics, 2026

    2026

  7. [14]

    Seeing further and wider: Joint spatio-temporal enlargement for micro-video popularity prediction, 2026

    Dali Wang, Yunyao Zhang, Junqing Yu, Yi-Ping Phoebe Chen, Chen Xu, and Zikai Song. Seeing further and wider: Joint spatio-temporal enlargement for micro-video popularity prediction, 2026. URLhttps://arxiv.org/abs/2604.20311

    2026 arXiv

  8. [15]

    Yunyao Zhang, Zikai Song, Hang Zhou, Wenfeng Ren, Yi-Ping Phoebe Chen, Junqing Yu, and Wei Yang.𝑔𝑎−𝑠 3: Comprehensive social network simulation with group agents. In Wanxiang Che, Joyce Nabende, Ekaterina Shutova, and Mohammad Taher Pilehvar, editors,Findings of the Associatio...

    2025 doi

  9. [16]

    Reliable reasoning: Learning and inference based on the ability of large language models

    Changsen Yuan, Rui Lin, and Cunhan Guo. Reliable reasoning: Learning and inference based on the ability of large language models. Applied Soft Computing, 190:114618, 2026. ISSN 1568-4946. doi: https://doi.org/10.1016/j.asoc.2026.114618. URLhttps://www. sciencedirect.com/scienc...

    2026

  10. [17]

    Linguistic complexity: Locality of syntactic dependencies.Cognition, 68(1):1–76, 1998

    Edward Gibson. Linguistic complexity: Locality of syntactic dependencies.Cognition, 68(1):1–76, 1998

    1998

  11. [18]

    Ambiguity, polysemy, and vagueness

    David Tuggy. Ambiguity, polysemy, and vagueness. 1993

    1993

  12. [19]

    Qingchuan Li, Jiatong Li, Zirui Liu, Mingyue Cheng, Yitong Zhou, Yuting Zeng, Qi Liu, and Tongxuan Liu. Are llms stable formal logic translators in logical reasoning across linguistically diversified texts? InProceedings of the ACM Web Conference 2026, WWW ’26, page 3633–3644,...

    2026

  13. [20]

    Cambridge University Press, 2003

    Peter Smith.An introduction to formal logic. Cambridge University Press, 2003

    2003

  14. [21]

    Chain-of-thought prompting elicits reasoning in large language models.Advances in neural information processing systems, 35:24824–24837, 2022

    Jason Wei, Xuezhi Wang, Dale Schuurmans, Maarten Bosma, Fei Xia, Ed Chi, Quoc V Le, Denny Zhou, et al. Chain-of-thought prompting elicits reasoning in large language models.Advances in neural information processing systems, 35:24824–24837, 2022

    2022

  15. [22]

    Divideandtranslate:Compositionalfirst-orderlogictranslationandverification for complex logical reasoning.arXiv preprint arXiv:2410.08047, 2024

    HyunRyu,GyeongmanKim,HyeminSLee,andEunhoYang. Divideandtranslate:Compositionalfirst-orderlogictranslationandverification for complex logical reasoning.arXiv preprint arXiv:2410.08047, 2024

    2024 arXiv

  16. [23]

    Cumulative reasoning with large language models.arXiv preprint arXiv:2308.04371, 2023

    Yifan Zhang, Jingqin Yang, Yang Yuan, and Andrew Chi-Chih Yao. Cumulative reasoning with large language models.arXiv preprint arXiv:2308.04371, 2023. Zhang et al.:Preprint submitted to ElsevierPage 18 of 19 Hexagonal Semantic Reasoning for LLMs

    2023 arXiv

  17. [24]

    Treeofthoughts:Deliberateproblem solving with large language models.Advances in neural information processing systems, 36:11809–11822, 2023

    ShunyuYao,DianYu,JeffreyZhao,IzhakShafran,TomGriffiths,YuanCao,andKarthikNarasimhan. Treeofthoughts:Deliberateproblem solving with large language models.Advances in neural information processing systems, 36:11809–11822, 2023

    2023

  18. [25]

    Determlr: Augmenting llm-based logical reasoning from indeterminacy to determinacy

    Hongda Sun, Weikai Xu, Wei Liu, Jian Luan, Bin Wang, Shuo Shang, Ji-Rong Wen, and Rui Yan. Determlr: Augmenting llm-based logical reasoning from indeterminacy to determinacy. InProceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: L...

    2024

  19. [26]

    Exploring the role of reasoning structures for constructing proofs in multi-step natural language reasoning with large language models.arXiv preprint arXiv:2410.08436, 2024

    Zi’ou Zheng, Christopher Malon, Martin Renqiang Min, and Xiaodan Zhu. Exploring the role of reasoning structures for constructing proofs in multi-step natural language reasoning with large language models.arXiv preprint arXiv:2410.08436, 2024

    arXiv

  20. [27]

    Harnessing the power of large language models for natural language to first-order logic translation.arXiv preprint arXiv:2305.15541, 2023

    Yuan Yang, Siheng Xiong, Ali Payani, Ehsan Shareghi, and Faramarz Fekri. Harnessing the power of large language models for natural language to first-order logic translation.arXiv preprint arXiv:2305.15541, 2023

    2023 arXiv

  21. [28]

    Aristotle: Mastering logical reasoning with a logic-complete decompose-search-resolve framework.arXiv preprint arXiv:2412.16953, 2024

    Jundong Xu, Hao Fei, Meng Luo, Qian Liu, Liangming Pan, William Yang Wang, Preslav Nakov, Mong-Li Lee, and Wynne Hsu. Aristotle: Mastering logical reasoning with a logic-complete decompose-search-resolve framework.arXiv preprint arXiv:2412.16953, 2024

    2024

  22. [29]

    Faithful logical reasoning via symbolic chain-of-thought

    Jundong Xu, Hao Fei, Liangming Pan, Qian Liu, Mong-Li Lee, and Wynne Hsu. Faithful logical reasoning via symbolic chain-of-thought. arXiv preprint arXiv:2405.18357, 2024

    2024 arXiv

  23. [30]

    Formal specifications from natural language.arXiv preprint arXiv:2206.01962, 2022

    Christopher Hahn, Frederik Schmitt, Julia J Tillman, Niklas Metzger, Julian Siber, and Bernd Finkbeiner. Formal specifications from natural language.arXiv preprint arXiv:2206.01962, 2022

    2022 arXiv

  24. [31]

    Logic-lm: Empowering large language models with symbolic solvers for faithful logical reasoning.arXiv preprint arXiv:2305.12295, 2023

    Liangming Pan, Alon Albalak, Xinyi Wang, and William Yang Wang. Logic-lm: Empowering large language models with symbolic solvers for faithful logical reasoning.arXiv preprint arXiv:2305.12295, 2023

    2023 arXiv

  25. [32]

    IndianaUniversity Press Bloomington, 1982

    AlgirdasJulienGreimas,JosephCourtés,LarryCrist,andDanielPatte.Semioticsandlanguage:Ananalyticaldictionary. IndianaUniversity Press Bloomington, 1982

    1982

  26. [33]

    On meaning: Selected writings in semiotic theory.(No Title), 1987

    Algirdas Julien Greimas. On meaning: Selected writings in semiotic theory.(No Title), 1987

    1987

  27. [34]

    Semantic-aware logical reasoning via a semiotic framework, 2026

    Yunyao Zhang, Xinglang Zhang, Junxi Sheng, Wenbing Li, Junqing Yu, Yi-Ping Phoebe Chen, Wei Yang, and Zikai Song. Semantic-aware logical reasoning via a semiotic framework, 2026

    2026

  28. [35]

    The power of the hexagon.Logica Universalis, 6(1):1–43, 2012

    Jean-Yves Béziau. The power of the hexagon.Logica Universalis, 6(1):1–43, 2012

    2012

  29. [36]

    Why the logical hexagon?Logica Universalis, 6(1):69–107, 2012

    Alessio Moretti. Why the logical hexagon?Logica Universalis, 6(1):69–107, 2012

    2012

  30. [37]

    From the logical square to blanché’s hexagon: formalization, applicability and the idea of the normative structure of thought.Logica Universalis, 6(1):45–67, 2012

    Aimable-André Dufatanye. From the logical square to blanché’s hexagon: formalization, applicability and the idea of the normative structure of thought.Logica Universalis, 6(1):45–67, 2012

    2012

  31. [38]

    The complexity of theorem-proving procedures

    Stephen A Cook. The complexity of theorem-proving procedures. InLogic, automata, and computational complexity: The works of Stephen A. Cook, pages 143–152. 2023

    2023

  32. [39]

    Mind the gap: Examining the self-improvement capabilities of large language models

    Yuda Song, Hanlin Zhang, Carson Eisenach, Sham Kakade, Dean Foster, and Udaya Ghai. Mind the gap: Examining the self-improvement capabilities of large language models. InInternational Conference on Learning Representations, volume 2025, pages 39894–39931, 2025

    2025

  33. [40]

    Folio: Natural language reasoning with first-order logic.arXiv preprint arXiv:2209.00840, 2022

    SimengHan,HaileySchoelkopf,YilunZhao,ZhentingQi,MartinRiddell,WenfeiZhou,JamesCoady,DavidPeng,YujieQiao,LukeBenson, et al. Folio: Natural language reasoning with first-order logic.arXiv preprint arXiv:2209.00840, 2022

    2022 arXiv

  34. [41]

    Language models are greedy reasoners: A systematic formal analysis of chain-of-thought.arXiv preprint arXiv:2210.01240, 2022

    Abulhair Saparov and He He. Language models are greedy reasoners: A systematic formal analysis of chain-of-thought.arXiv preprint arXiv:2210.01240, 2022

    2022 arXiv

  35. [42]

    Proofwriter: Generating implications, proofs, and abductive statements over natural language.arXiv preprint arXiv:2012.13048, 2020

    Oyvind Tafjord, Bhavana Dalvi Mishra, and Peter Clark. Proofwriter: Generating implications, proofs, and abductive statements over natural language.arXiv preprint arXiv:2012.13048, 2020

    2012 arXiv

  36. [43]

    Largelanguagemodelsmeetsymbolic provers for logical reasoning evaluation.arXiv preprint arXiv:2502.06563, 2025

    ChengwenQi,RenMa,BowenLi,HeDu,BinyuanHui,JinwangWu,YuanjunLaili,andConghuiHe. Largelanguagemodelsmeetsymbolic provers for logical reasoning evaluation.arXiv preprint arXiv:2502.06563, 2025

    2025 arXiv

  37. [45]

    Coupled mamba: Enhanced multi-modal fusion with coupled state space model.arXiv preprint arXiv:2405.18014, 2024

    Wenbing Li, Hang Zhou, Junqing Yu, Zikai Song, and Wei Yang. Coupled mamba: Enhanced multi-modal fusion with coupled state space model.arXiv preprint arXiv:2405.18014, 2024

    2024 arXiv

  38. [46]

    Lora-mixer:Coordinatemodularloraexpertsthroughserial attention routing.arXiv preprint arXiv:2507.00029, 2025

    WenbingLi,ZikaiSong,HangZhou,YunyaoZhang,JunqingYu,andWeiYang. Lora-mixer:Coordinatemodularloraexpertsthroughserial attention routing.arXiv preprint arXiv:2507.00029, 2025

    2025 arXiv

  39. [47]

    Hotcomment: A benchmark for evaluating popularity of online comments.arXiv preprint arXiv:2604.25614, 2026

    Yafeng Wu, Yunyao Zhang, Liliang Ye, Guiyi Zeng, Junqing Yu, Chen Xu, and Zikai Song. Hotcomment: A benchmark for evaluating popularity of online comments.arXiv preprint arXiv:2604.25614, 2026

    2026 arXiv

  40. [48]

    Self-consistency improves chain of thought reasoning in language models.arXiv preprint arXiv:2203.11171, 2022

    Xuezhi Wang, Jason Wei, Dale Schuurmans, Quoc Le, Ed Chi, Sharan Narang, Aakanksha Chowdhery, and Denny Zhou. Self-consistency improves chain of thought reasoning in language models.arXiv preprint arXiv:2203.11171, 2022

    2022 arXiv

  41. [49]

    Whypromptdesignmattersandworks:Acomplexityanalysisofprompt search space in llms.arXiv preprint arXiv:2503.10084, 2025

    XiangZhang,JuntaiCao,JiaqiWei,ChenyuYou,andDujianDing. Whypromptdesignmattersandworks:Acomplexityanalysisofprompt search space in llms.arXiv preprint arXiv:2503.10084, 2025

    2025 arXiv

  42. [50]

    Gsm-symbolic: Understanding the limitations of mathematical reasoning in large language models.arXiv preprint arXiv:2410.05229, 2024

    Iman Mirzadeh, Keivan Alizadeh, Hooman Shahrokhi, Oncel Tuzel, Samy Bengio, and Mehrdad Farajtabar. Gsm-symbolic: Understanding the limitations of mathematical reasoning in large language models.arXiv preprint arXiv:2410.05229, 2024

    arXiv

  43. [51]

    Efficientreasoningmodels:Asurvey.arXivpreprintarXiv:2504.10903,2025

    SichengFeng,GongfanFang,XinyinMa,andXinchaoWang. Efficientreasoningmodels:Asurvey.arXivpreprintarXiv:2504.10903,2025

    2025

  44. [52]

    Magic-vqa: Multimodal and grounded inference with commonsense knowledge for visual question answering

    Shuo Yang, Caren Han, Siwen Luo, and Eduard Hovy. Magic-vqa: Multimodal and grounded inference with commonsense knowledge for visual question answering. InFindings of the Association for Computational Linguistics: ACL 2025, pages 16967–16986, 2025

    2025

  45. [53]

    Can mllms guide me home? a benchmark study on fine-grained visual reasoning from transit maps.arXiv preprint arXiv:2505.18675, 2025

    Sicheng Feng, Song Wang, Shuyi Ouyang, Lingdong Kong, Zikai Song, Jianke Zhu, Huan Wang, and Xinchao Wang. Can mllms guide me home? a benchmark study on fine-grained visual reasoning from transit maps.arXiv preprint arXiv:2505.18675, 2025

    2025

  46. [54]

    Double:Breakingtheaccelerationlimitviadouble retrieval speculative parallelism, 2026

    YuhaoShen,TianyuLiu,JunyiShen,JinyangWu,QuanKong,LiHuan,andCongWang. Double:Breakingtheaccelerationlimitviadouble retrieval speculative parallelism, 2026

    2026

  47. [55]

    Explicit planning helps language models in logical reasoning.arXiv preprint arXiv:2303.15714, 2023

    Hongyu Zhao, Kangrui Wang, Mo Yu, and Hongyuan Mei. Explicit planning helps language models in logical reasoning.arXiv preprint arXiv:2303.15714, 2023

    2023 arXiv

  48. [56]

    Logic and semantic analysis

    Ernest Lepore and Matthew Stone. Logic and semantic analysis. InPhilosophy of Logic, pages 173–204. Elsevier, 2007

    2007

  49. [57]

    Maupassant: The semiotics of text

    Algirdas Julien Greimas. Maupassant: The semiotics of text. 1988. Zhang et al.:Preprint submitted to ElsevierPage 19 of 19

    1988

Pith tools

Reviewed August 15, 2026 · model on record in the stance chip above.