mlsanityData to 14 Oct 2026

The Thin Line Between Comprehension and Persuasion in LLMs

arXiv 2507.01936 ↗First posted 2 Jul 2025cs.CL cs.CYcited by 5

Not on the map. Why?

Abstract (arXiv)

Large language models (LLMs) are excellent at maintaining high-level, convincing dialogue, but it remains unclear whether their persuasive success reflects genuine understanding of the discourse. We examine this question through informal debates between humans and LLMs, first by measuring their persuasive skills, and then by relating these to their understanding of _what_ is being talked about: namely, their comprehension of argumentative structures and the pragmatic context on the same debates. We find that LLMs effectively maintain coherent, persuasive debates, and can sway the beliefs of both participants and audiences. We also note that awareness or suspicion of AI involvement encourage people to be more critical of the arguments made. However, we also find that LLMs are unable to show comprehension of deeper dialogical structures, such as argument quality or existence of supporting premises. Our results reveal a disconnect between LLM comprehension and dialogical skills, raising ethical and practical concerns on their deployment on explanation-critical contexts. From an argumentation-theoretical perspective, we experimentally question whether an agent, if it can convincingly maintain a dialogue, is required to show it knows what is talking about.

Key facts. Cited by 5 papers in the map (data through 14 Oct 2026). Builds on 57 of its 108 references that are in the map. First posted on arXiv 2 Jul 2025.

Builds on

57 of 108 references.

References · 57

  1. 2023Judging LLM-as-a-Judge with MT-Bench and Chatbot Arenacited by 8,626
  2. 2013Intriguing properties of neural networkscited by 5,155
  3. 2023G-Eval: NLG Evaluation using GPT-4 with Better Human Alignmentcited by 2,276
  4. 2024A Survey on LLM-as-a-Judgecited by 1,423
  5. 2024Chatbot Arena: An Open Platform for Evaluating LLMs by Human Preferencecited by 1,184
  6. 2021Fantastically Ordered Prompts and Where to Find Them: Overcoming Few-Shot Prompt Order Sensitivitycited by 1,135
  7. 2023Language Models Don't Always Say What They Think: Unfaithful Explanations in Chain-of-Thought Promptingcited by 1,122
  8. 2021Probing Classifiers: Promises, Shortcomings, and Advancescited by 876
  9. 2023Can Large Language Models Be an Alternative to Human Evaluations?cited by 734
  10. 2021BARTScore: Evaluating Generated Text as Text Generationcited by 722
  11. 2023Automatic Prompt Optimization with "Gradient Descent" and Beam Searchcited by 653
  12. 2023Are Emergent Abilities of Large Language Models a Mirage?cited by 523
38 more
  1. 2023Faith and Fate: Limits of Transformers on Compositionalitycited by 503
  2. 2024LLMs-as-Judges: A Comprehensive Survey on LLM-based Evaluation Methodscited by 465
  3. 2023NLP Evaluation in trouble: On the Need to Measure LLM Data Contamination for each Benchmarkcited by 348
  4. 2021Do Prompt-Based Models Really Understand the Meaning of their Prompts?cited by 326
  5. 2024Humans or LLMs as the Judge? A Study on Judgement Biasescited by 323
  6. 2023Large Language Models Fail on Trivial Alterations to Theory-of-Mind Taskscited by 292
  7. 2021All That's 'Human' Is Not Gold: Evaluating Human Evaluation of Generated Textcited by 287
  8. 2020Evaluation of Text Generation: A Surveycited by 277
  9. 2024Debating with More Persuasive LLMs Leads to More Truthful Answerscited by 266
  10. 2022Reasoning with Language Model Prompting: A Surveycited by 252
  11. 2024LLMs instead of Human Judges? A Large Scale Empirical Study across 20 NLP Evaluation Taskscited by 243
  12. 2020Information-Theoretic Probing for Linguistic Structurecited by 207
  13. 2023Do LLMs exhibit human-like response biases? A case study in survey designcited by 180
  14. 2022The Unreliability of Explanations in Few-shot Prompting for Textual Reasoningcited by 169
  15. 2022Repairing the Cracked Foundation: A Survey of Obstacles in Evaluation Practices for Generated Textcited by 155
  16. 2025MME-CoT: Benchmarking Chain-of-Thought in Large Multimodal Models for Reasoning Quality, Robustness, and Efficiencycited by 125
  17. 2019Learning to Deceive with Attention-Based Explanationscited by 111
  18. 2021Transformer visualization via dictionary learning: contextualized embedding as a linear superposition of transformer factorscited by 108
  19. 2021An Interpretability Illusion for BERTcited by 97
  20. 2023Are Large Language Model-based Evaluators the Solution to Scaling Up Multilingual Evaluation?cited by 88
  21. 2023LLMs and the Abstraction and Reasoning Corpus: Successes, Failures, and the Importance of Object-based Representationscited by 82
  22. 2022Do Language Models Plagiarize?cited by 73
  23. 2024A Chain-of-Thought Is as Strong as Its Weakest Link: A Benchmark for Verifiers of Reasoning Chainscited by 65
  24. 2024SimpleToM: Exposing the Gap between Explicit ToM Inference and Implicit ToM Application in LLMscited by 57
  25. 2018Dialogue Learning with Human Teaching and Feedback in End-to-End Trainable Task-Oriented Dialogue Systemscited by 53
  26. 2023Theory of Mind in Large Language Models: Examining Performance of 11 State-of-the-Art models vs. Children Aged 7-10 on Advanced Testscited by 48
  27. 2024Understanding and Patching Compositional Reasoning in LLMscited by 46
  28. 2024Cheating Automatic LLM Benchmarks: Null Models Achieve High Win Ratescited by 45
  29. 2023Exploring the Potential of Large Language Models in Computational Argumentationcited by 42
  30. 2025MegaTTS 3: Sparse Alignment Enhanced Latent Diffusion Transformer for Zero-Shot Speech Synthesiscited by 41
  31. 2024Social Bias Evaluation for Large Language Models Requires Prompt Variationscited by 40
  32. 2024Do Large Language Models Have Compositional Ability? An Investigation into Limitations and Scalabilitycited by 40
  33. 2024Human-Centered Design Recommendations for LLM-as-a-Judgecited by 37
  34. 2024RTP-LX: Can LLMs Evaluate Toxicity in Multilingual Scenarios?cited by 36
  35. 2024Mathematical Models of Computation in Superpositioncited by 35
  36. 2025ThinkBench: Dynamic Out-of-Distribution Evaluation for Robust LLM Reasoningcited by 17
  37. 2024Exploring the Compositional Deficiency of Large Language Models in Mathematical Reasoningcited by 15
  38. 2024ArgMed-Agents: Explainable Clinical Decision Reasoning with LLM Disscusion via Argumentation Schemescited by 13

50 most cited of 57 All 57 (JSON)

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 2; 2026: 3.

5 papers.

  1. 2025Democratizing Diplomacy: A Harness for Evaluating Any Large Language Model on Full-Press Diplomacycited by 7
  2. 2025Is In-Context Learning Learning?cited by 2
  3. 2026If LLMs Have Human-Like Attributes, Then So Does Age of Empires IIcited by 2
  4. 2026MA2P: A Meta-Cognitive Autonomous Intelligent Agents Framework for Complex Persuasioncited by 0
  5. 2026Evaluating the Capabilities of LLMs for Persuasive Dialoguecited by 0