Elamur Research⌕
☰
home / papers / 2026 / Argument-Based Consistency in Toxicity Explanations of LLMs

Argument-Based Consistency in Toxicity Explanations of LLMs

Paper · 2026 · Ramaravind Kommiya Mothilal, Joanna Roy, Syed Ishtiaque Ahmed et al.

Summary

Most NLP work on toxicity and LLMs focuses on detection. This paper asks instead how well models reason when they explain their toxicity judgments in free-form text. Existing explainability criteria rely on input perturbations or label shifts, so they fit such explanations poorly. Drawing on Informal Logic, the authors treat an explanation as an argument backing a stance. They propose Argument-based Consistency (ArC), a criterion with five dimensions:

  • relevance (REL)
  • internal reliance (INT)
  • external reliance (EXT)
  • individual sufficiency (SUF)
  • individual necessity (NEC)

ArC is measured with six metrics based on uncertainty quantification. A three-stage prompting pipeline (JUSTIFY, UPHOLD-REASONS, UPHOLD-STANCE) was applied to three Llama models and Ministral-8B across five toxicity datasets.

Models give plausible justifications at first. They falter, however, when asked how individual reasons relate to the full reason set and to the stance. Responses then become inconsistent or off-topic, and larger size does not reliably help. The work argues for evaluating toxicity reasoning, not just detection, in socially sensitive applications. Code and generated explanations are released.

Key points

  • Logic behind SUF and NEC:
    • For a toxic stance, any single reason should suffice, since one violation is enough to cause harm.
    • For a non-toxic stance, every reason should be necessary, since all of them together establish safety.
  • Models and resources:
    • Models evaluated: Llama-3B, Llama-8B, Llama-70B and Ministral-8B.
    • Released resource: over 20K LLM-generated multi-reason toxicity explanations.
  • Model behaviour:
    • Llama-70B produced the most diverse relevant reasons.
    • It was the only model that largely upheld its own stated reasons.
    • Even so, it still gave inconsistent responses on the SUF and NEC dimensions.
  • NEC scores:
    • Ministral-8B scored highest on NEC, averaging 0.297.
    • The authors attribute this to label bias: the model tends to flag insufficiency regardless of stance, rather than genuinely reasoning about necessity.
  • Limitations:
    • The metrics need token logits, so they do not apply to black-box LLMs.
    • They depend on semantic similarity.
    • There is no meta-evaluation against a baseline or human annotations.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

 Fields, topics and concepts

📚 Shelves 6The field of science this work belongs to (OpenAlex fields).Computer Science · 18,614 worksArtificial Intelligence · 17,965 worksSocial Sciences · 1,954 worksSafety Research · 626 worksEngineering · 188 worksSafety, Risk, Reliability and Quality · 14 works

🧭 Topics 2Research subjects and languages the work is about.Language models & pretraining · 3,894 worksBias, fairness & safety · 587 works

Cite

Cite as Ramaravind Kommiya Mothilal, Joanna Roy, Syed Ishtiaque Ahmed and Shion Guha. 2026. Argument-Based Consistency in Toxicity Explanations of LLMs. Findings of the Association for Computational Linguistics: EACL 2026.

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2026.findings-eacl.310

@inproceedings{mothilal2026argument,
  title = {{Argument-Based Consistency in Toxicity Explanations of LLMs}},
  author = {Ramaravind Kommiya Mothilal, Joanna Roy, Syed Ishtiaque Ahmed and Shion Guha},
  booktitle = {Findings of the Association for Computational Linguistics: EACL 2026},
  year = {2026},
  doi = {10.18653/v1/2026.findings-eacl.310},
  url = {https://doi.org/10.18653/v1/2026.findings-eacl.310}
}

References

104 references, 28 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Chirag Agarwal, Sree Harsha Tanneru, and Himabindu Lakkaraju. 2024. Faithfulness vs. plausibility: On the (un) reliability of explanations from large language models. arXiv preprint arXiv:2402.04614. — [arXiv:2402.04614](https://arxiv.org/abs/2402.04614) find
  2. Blair J Anthony and Informal Logic. 2004. Argument and its uses. Informal Logic, 24(2):137–151. find
  3. Lora Aroyo, Alex Taylor, Mark Diaz, Christopher Homan, Alicia Parrish, Gregory Serapio-García, Vinodkumar Prabhakaran, and Ding Wang. 2023. Dices dataset: Diversity in conversational ai evaluation for safety. Advances in Neural Information Processing Systems, 36:53330–53342. find
  4. Pepa Atanasova, Oana-Maria Camburu, Christina Lioma, Thomas Lukasiewicz, Jakob Grue Simonsen, and Isabelle Augenstein. 2023. Faithfulness tests for natural language explanations. In The 61st Annual Meeting Of The Association For Computational Linguistics.
  5. Robert Audi and 1 others. 1999. The Cambridge dictionary of philosophy, volume 2. Cambridge university press Cambridge. find
  6. Esma Balkir, Isar Nejadgholi, Kathleen C Fraser, and Svetlana Kiritchenko. 2022. Necessity and sufficiency for explaining text classifiers: A case study in hate speech detection. In Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Hum… find
  7. Else M Barth and Erik C Krabbe. 2010. From axiom to dialogue: A philosophical study of logics and argumentation. Walter de Gruyter. find
  8. Federico Bianchi, Stefanie Anja Hills, Patricia Rossini, Dirk Hovy, Rebekah Tromble, and Nava Tintarev. 2022. " it’s not just hate”: A multi-dimensional perspective on detecting harmful speech online. arXiv preprint arXiv:2210.15870. find
  9. J Anthony Blair. 2019. Judging arguments. Studies in Critical Thinking, pages 225–247. find
  10. J Anthony Blair, Derek Allen, Sharon Bailin, Ashley Barnett, Mark Battersby, Yiwen Dai, Martin Davies, Robert H Ennis, Alec Fisher, Tim van Gelder, and 1 others. 2021. Studies in Critical Thinking. Windsor Studies in Argumentation. find
  11. J Anthony Blair and J Anthony Blair. 2012. Informal logic and logic. Groundwork in the Theory of Argumentation: Selected Papers of J. Anthony Blair, pages 119–132. find
  12. Daniel Borkan, Lucas Dixon, Jeffrey Sorensen, Nithum Thain, and Lucy Vasserman. 2019. Nuanced metrics for measuring unintended bias with real data for text classification. In Companion proceedings of the 2019 world wide web conference, pages 491–500. find
  13. Xi Chen and Shuo Wang. 2025. Pragmatic Inference Chain (PIC) improving LLMs’ reasoning of authentic implicit toxic language. arXiv [cs.CL].
  14. Yanda Chen, Ruiqi Zhong, Narutatsu Ri, Chen Zhao, He He, Jacob Steinhardt, Zhou Yu, and Kathleen McKeown. 2023. Do models explain themselves? counterfactual simulatability of natural language explanations. arXiv preprint arXiv:2307.08678. find
  15. Gao Chujie, Siyuan Wu, Yue Huang, Dongping Chen, Qihui Zhang, Zhengyan Fu, Yao Wan, Lichao Sun, and Xiangliang Zhang. 2024. Honestllm: Toward an honest and helpful large language model. Advances in Neural Information Processing Systems, 37:7213– 7255. find
  16. Sanjoy Dasgupta, Nave Frost, and Michal Moshkovitz. 2022. Framework for evaluating faithfulness of local explanations. In International Conference on Machine Learning, pages 4794–4815. PMLR. find
  17. Jay DeYoung, Sarthak Jain, Nazneen Fatema Rajani, Eric Lehman, Caiming Xiong, Richard Socher, and Byron C Wallace. 2020. Eraser: A benchmark to evaluate rationalized nlp models. In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics, pages 4443–4458.
  18. Jinhao Duan, Hao Cheng, Shiqi Wang, Alex Zavalny, Chenan Wang, Renjing Xu, Bhavya Kailkhura, and Kaidi Xu. 2023. Shifting attention to relevance: Towards the uncertainty estimation of large language models. find
  19. Robert H Ennis. 1996. Critical thinking upper saddle river. NJ: Practice Hall. find
  20. Ekaterina Fadeeva, Aleksandr Rubashevskii, Artem Shelmanov, Sergey Petrakov, Haonan Li, Hamdy Mubarak, Evgenii Tsymbalov, Gleb Kuzmin, Alexander Panchenko, Timothy Baldwin, and 1 others. 2024. Fact-checking the output of large language models via token-level uncertainty quantification. arXiv preprin…
  21. Mohsen Fayyaz, Fan Yin, Jiao Sun, and Nanyun Peng. 2024. Evaluating human alignment and model faithfulness of llm rationale. arXiv preprint arXiv:2407.00219. — [arXiv:2407.00219](https://arxiv.org/abs/2407.00219) find
  22. Robert J. Fogelin. 1991. Understanding Arguments: An Introduction to Informal Logic. Harcourt Brace Jovanovich, San Diego. find
  23. Paula Fortuna, Juan Soler, and Leo Wanner. 2020. Toxic, hateful, offensive or abusive? what are we really classifying? an empirical analysis of hate speech datasets. In Proceedings of the Twelfth Language Resources and Evaluation Conference, pages 6786– 6794.
  24. Samuel Gehman, Suchin Gururangan, Maarten Sap, Yejin Choi, and Noah A Smith. 2020. Realtoxicityprompts: Evaluating neural toxic degeneration in language models. In Findings of the Association for Computational Linguistics: EMNLP 2020, pages 3356–3369.
  25. Aaron Gokaslan and Vanya Cohen. 2019. Openwebtext corpus. http://Skylion007.github.io/ OpenWebTextCorpus. ↗
  26. Trudy Govier and 1 others. 2010. A practical study of argument. Wadsworth Belmont, CA. find
  27. Nitesh Goyal, Ian D Kivlichan, Rachel Rosen, and Lucy Vasserman. 2022. Is your toxicity my toxicity? exploring the impact of rater identity on toxicity annotation. Proceedings of the ACM on Human-Computer Interaction, 6(CSCW2):1–28. find
  28. Robert Grootendorst, Frans H Van Eemeren, and Rob Grootendorst. 2004. A systematic theory of argumentation: The pragma-dialectical approach. Cambridge University Press. find
  29. Simeng Han, Hailey Schoelkopf, Yilun Zhao, Zhenting Qi, Martin Riddell, Wenfei Zhou, James Coady, David Peng, Yujie Qiao, Luke Benson, and 1 others. 2024. Folio: Natural language reasoning with first-order logic. In Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processi…
  30. Gilbert Harman. 1984. Logic and reasoning. In Foundations: logic, language, and mathematics, pages 107–127. Springer. find
  31. Thomas Hartvigsen, Saadia Gabriel, Hamid Palangi, Maarten Sap, Dipankar Ray, and Ece Kamar. 2022. Toxigen: A large-scale machine-generated dataset for adversarial and implicit hate speech detection. In Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1:…
  32. Xinlei He, Savvas Zannettou, Yun Shen, and Yang Zhang. 2023. You only prompt once: On the capabilities of prompt learning on large language models to tackle toxic content. arXiv [cs.CL]. find
  33. David Hitchcock. 2007. Informal logic and the concept of argument. In Philosophy of logic, pages 101–129. Elsevier. find
  34. David Hitchcock. 2021. The concept of an argument. Studies in Critical Thinking. find
  35. Alon Jacovi and Yoav Goldberg. 2020. Towards faithfully interpretable nlp systems: How should we define and evaluate faithfulness? arXiv preprint arXiv:2004.03685.
  36. Alon Jacovi and Yoav Goldberg. 2021. Aligning faithful interpretations with their social attribution. Transactions of the Association for Computational Linguistics, 9:294–310.
  37. Zhengbao Jiang, Jun Araki, Haibo Ding, and Graham Neubig. 2021. How can we know when language models know? on the calibration of language models for question answering. Transactions of the Association for Computational Linguistics, 9:962–977. find
  38. Ralph H Johnson. 1996. The need for a dialectical tier in arguments. In International Conference on Formal and Applied Practical Reasoning, pages 349–360. Springer. find
  39. Ralph H Johnson. 1999. The relation between formal and informal logic. Argumentation, 13:265–274. find
  40. Ralph H Johnson. 2009. Some reflections on the informal logic initiative. Studies in logic, grammar and rhetoric, 16(29):17–46. find
  41. Ralph H Johnson and J Anthony Blair. 1994. Informal logic: Past and present. New essays in informal logic, pages 1–19. find
  42. Ralph H Johnson and J Anthony Blair. 2002. Informal logic and the reconfiguration of logic. In Studies in Logic and Practical Reasoning, volume 1, pages 339–396. Elsevier. find
  43. Ralph Henry Johnson and J Anthony Blair. 2006. Logical self-defense. Idea. find
  44. Md Tawkat Islam Khondaker, Muhammad Abdul-Mageed, and Laks VS Lakshmanan. 2024. Detoxllm: A framework for detoxification with explanations. arXiv preprint arXiv:2402.15951. find
  45. Pieter-Jan Kindermans, Sara Hooker, Julius Adebayo, Maximilian Alber, Kristof T Schütt, Sven Dähne, Dumitru Erhan, and Been Kim. 2019. The (un) reliability of saliency methods. Explainable AI: Interpreting, explaining and visualizing deep learning, pages 267–280. find
  46. Xiang Lisa Li, Urvashi Khandelwal, and Kelvin Guu. 2024. Few-shot recalibration of language models. arXiv preprint arXiv:2403.18286. — [arXiv:2403.18286](https://arxiv.org/abs/2403.18286) find
  47. Zhen Lin, Shubhendu Trivedi, and Jimeng Sun. 2023. Generating with confidence: Uncertainty quantification for black-box large language models. arXiv preprint arXiv:2305.19187. find
  48. Xiaoou Liu, Tiejin Chen, Longchao Da, Chacha Chen, Zhen Lin, and Hua Wei. 2025. Uncertainty quantification and confidence calibration in large language models: A survey. arXiv preprint arXiv:2503.15850. — [arXiv:2503.15850](https://arxiv.org/abs/2503.15850)
  49. Yibing Liu, Haoliang Li, Yangyang Guo, Chenqi Kong, Jing Li, and Shiqi Wang. 2022. Rethinking attention-model explainability through faithfulness violation test. In International conference on machine learning, pages 13807–13824. PMLR. find
  50. Yinhan Liu, Myle Ott, Naman Goyal, Jingfei Du, Mandar Joshi, Danqi Chen, Omer Levy, Mike Lewis, Luke Zettlemoyer, and Veselin Stoyanov. 2019. Roberta: A robustly optimized bert pretraining approach. arXiv preprint arXiv:1907.11692.
  51. Chengda Lu, Xiaoyu Fan, Yu Huang, Rongwu Xu, Jijie Li, and Wei Xu. 2025. Does chain-of-thought reasoning really reduce harmfulness from jailbreaking? arXiv preprint arXiv:2505.17650. find
  52. Scott M Lundberg and Su-In Lee. 2017. A unified approach to interpreting model predictions. Advances in neural information processing systems, 30.
  53. Qing Lyu, Marianna Apidianaki, and Chris Callison-Burch. 2024. Towards faithful model explanation in nlp: A survey. Computational Linguistics, 50(2):657– 723.
  54. Potsawee Manakul, Adian Liusie, and Mark JF Gales. 2023. Selfcheckgpt: Zero-resource black-box hallucination detection for generative large language models. arXiv preprint arXiv:2303.08896. find
  55. Binny Mathew, Punyajoy Saha, Seid Muhie Yimam, Chris Biemann, Pawan Goyal, and Animesh Mukherjee. 2020. HateXplain: A benchmark dataset for explainable hate speech detection. arXiv [cs.CL]. find
  56. Hyukhun Koh, Dohyung Kim, Minwoo Lee, and Kyomin Jung. 2024. Can LLMs recognize toxicity? A structured investigation framework and toxicity metric. arXiv [cs.CL]. find
  57. Katie Matton, Robert Osazuwa Ness, John Guttag, and Emre Kıcıman. 2025. Walk the talk? measuring the faithfulness of large language model explanations. arXiv preprint arXiv:2504.14150. — [arXiv:2504.14150](https://arxiv.org/abs/2504.14150) find
  58. Tamera Lanham, Anna Chen, Ansh Radhakrishnan, Benoit Steiner, Carson Denison, Danny Hernandez, Dustin Li, Esin Durmus, Evan Hubinger, Jackson Kernion, and 1 others. 2023. Measuring faithfulness in chain-of-thought reasoning. arXiv preprint arXiv:2307.13702.
  59. Ramaravind Kommiya Mothilal, Faisal M Lalani, Syed Ishtiaque Ahmed, Shion Guha, and Sharifa Sultana. 2025. Talking about the assumption in the room. In Proceedings of the 2025 CHI Conference on Human Factors in Computing Systems, pages 1–16. find
  60. Nayeon Lee, Chani Jung, and Alice Oh. 2023. Hate speech classifiers are culturally insensitive. In Proceedings of the First Workshop on Cross-Cultural Considerations in NLP (C3NLP), pages 35–46. find
  61. Meike Nauta, Jan Trienes, Shreyasi Pathak, Elisa Nguyen, Michelle Peters, Yasmin Schmitt, Jörg Schlötterer, Maurice Van Keulen, and Christin Seifert. 2023. From anecdotal evidence to quantitative evaluation methods: A systematic review on evaluating explainable ai. ACM Computing Surveys, 55(13s):1–… find
  62. Ayushi Nirmal, Amrita Bhattacharjee, and Paras Sheth Huan Liu. 2024. Towards interpretable hate speech detection using large language model-extracted rationales. In The 8th Workshop on Online Abuse and Harms (WOAH), page 223. find
  63. Daniel J O’Keefe. 1977. Two concepts of argument. The Journal of the American Forensic Association, 13(3):121–128. find
  64. Letitia Parcalabescu and Anette Frank. 2023. On measuring faithfulness or self-consistency of natural language explanations. arXiv preprint arXiv:2311.07466. find
  65. Alicia Parrish, Vinodkumar Prabhakaran, Lora Aroyo, Mark Díaz, Christopher M Homan, Greg Serapio-García, Alex S Taylor, and Ding Wang. 2024. Diversity-aware annotation for conversational ai safety. In Proceedings of Safety4ConvAI: The Third Workshop on Safety for Conversational AI@ LREC-COLING 2024,… find
  66. Archiki Prasad, Swarnadeep Saha, Xiang Zhou, and Mohit Bansal. 2023. Receval: Evaluating reasoning chains via correctness and informativeness. arXiv preprint arXiv:2304.10703.
  67. Ansh Radhakrishnan, Karina Nguyen, Anna Chen, Carol Chen, Carson Denison, Danny Hernandez, Esin Durmus, Evan Hubinger, Jackson Kernion, Kamilė Lukošiūtė, and 1 others. 2023. Question decomposition improves the faithfulness of model-generated reasoning. arXiv preprint arXiv:2307.11768. find
  68. Alan Ramponi and Sara Tonelli. 2022. Features or spurious artifacts? data-centric baselines for fair and robust hate speech detection. In Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 3027–3040. A… find
  69. Nils Reimers and Iryna Gurevych. 2019. Sentence-bert: Sentence embeddings using siamese bert-networks. In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing. Association for Computational Linguistics.
  70. Maarten Sap, Dallas Card, Saadia Gabriel, Yejin Choi, and Noah A Smith. 2019. The risk of racial bias in hate speech detection. In Proceedings of the 57th annual meeting of the association for computational linguistics, pages 1668–1678.
  71. Maarten Sap, Swabha Swayamdipta, Laura Vianna, Xuhui Zhou, Yejin Choi, and Noah A Smith. 2022. Annotators with Attitudes: How Annotator Beliefs And Identities Bias Toxic Language Detection. In Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Lingu…
  72. Thomas Savage, John Wang, Robert Gallo, Abdessalem Boukil, Vishwesh Patel, Seyed Amir Ahmad Safavi-Naini, Ali Soroush, and Jonathan H Chen. 2024. Large language model uncertainty measurement and calibration for medical diagnosis and treatment. medRxiv, pages 2024–06. find
  73. Michael Scriven. 1980. The philosophical and pragmatic significance of informal logic. In Informal logic: The first international symposium. find
  74. Omar Shaikh, Hongxin Zhang, William Held, Michael Bernstein, and Diyi Yang. 2023. On second thought, let’s not think step by step! bias and toxicity in zero-shot reasoning. In Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 4454–…
  75. Yaozong Shen, Lijie Wang, Ying Chen, Xinyan Xiao, Jing Liu, and Hua Wu. 2022. An interpretability evaluation benchmark for pre-trained language models. arXiv preprint arXiv:2207.13948. — [arXiv:2207.13948](https://arxiv.org/abs/2207.13948) find
  76. Suzanna Sia, Anton Belyy, Amjad Almahairi, Madian Khabsa, Luke Zettlemoyer, and Lambert Mathias. 2023. Logical satisfiability of counterfactuals for faithful explanations in nli. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 37, pages 9837–9845. find
  77. Noah Siegel, Oana-Maria Camburu, Nicolas Heess, and Maria Perez-Ortiz. 2024. The probabilities also matter: A more faithful metric for faithfulness of free-text explanations in large language models. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 2… find
  78. Aviv Slobodkin, Omer Goldman, Avi Caciularu, Ido Dagan, and Shauli Ravfogel. 2023. The curious case of hallucinatory (un) answerability: Finding truths in the hidden states of over-confident large language models. arXiv preprint arXiv:2310.11877. find
  79. Kaya Stechly, Karthik Valmeekam, and Subbarao Kambhampati. 2024. Chain of thoughtlessness? an analysis of cot in planning. Advances in Neural Information Processing Systems, 37:29106–29141. find
  80. Mukund Sundararajan, Ankur Taly, and Qiqi Yan. 2017. Axiomatic attribution for deep networks. In International conference on machine learning, pages 3319– 3328. PMLR. find
  81. Jidong Tian, Yitian Li, Wenqing Chen, Liqiang Xiao, Hao He, and Yaohui Jin. 2021. Diagnosing the first-order logical reasoning ability through logicnli. In Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pages 3738–3747.
  82. Stephen E Toulmin. 1992. Logic, rhetoric, and reason: Redressing the balance. Argumentation illuminated, pages 3–11. find
  83. Stephen E Toulmin. 2003. The uses of argument. Cambridge university press. find
  84. Yi-Lin Tuan, Xilun Chen, Eric Michael Smith, Louis Martin, Soumya Batra, Asli Celikyilmaz, William Yang Wang, and Daniel M Bikel. 2024. Towards safety and helpfulness balanced responses via controllable large language models. arXiv preprint arXiv:2404.01295. — [arXiv:2404.01295](https://arxiv.org/abs/2404.01295) find
  85. Miles Turpin, Julian Michael, Ethan Perez, and Samuel Bowman. 2023. Language models don’t always say what they think: Unfaithful explanations in chain-of-thought prompting. Advances in Neural Information Processing Systems, 36:74952–74965.
  86. Dirk Van Dalen. 2004. Logic and structure. Springer. find
  87. Guillermo Villate-Castillo, Javier Del Ser, and Borja Sanz Urquijo. 2024. A systematic review of toxicity in Large Language Models: Definitions, datasets, detectors, detoxification methods and challenges. Research Square. find
  88. Douglas Walton. 2013. Argumentation schemes for presumptive reasoning. Routledge. find
  89. Douglas N Walton. 1990. What is reasoning? what is an argument? The journal of Philosophy, 87(8):399– 419. find
  90. Ruida Wang, Yuxin Li, Yi R Fung, and Tong Zhang. 2025. Let’s reason formally: Natural-formal hybrid reasoning enhances llm’s math capability. arXiv preprint arXiv:2505.23703.
  91. Zhiyuan Wang, Jinhao Duan, Lu Cheng, Yue Zhang, Qingni Wang, Xiaoshuang Shi, Kaidi Xu, Hengtao Shen, and Xiaofeng Zhu. 2024. Conu: Conformal uncertainty in large language models with correctness coverage guarantees. arXiv preprint arXiv:2407.00499. find
  92. Jason Wei, Xuezhi Wang, Dale Schuurmans, Maarten Bosma, Fei Xia, Ed Chi, Quoc V Le, Denny Zhou, and 1 others. 2022. Chain-of-thought prompting elicits reasoning in large language models. Advances in neural information processing systems, 35:24824– 24837.
  93. Jiaxin Wen, Pei Ke, Hao Sun, Zhexin Zhang, Chengfei Li, Jinfeng Bai, and Minlie Huang. 2023. Unveiling the implicit toxicity in large language models. In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, pages 1322–1338.
  94. Joseph W Wenzel, William L Benoit, Dale Hample, and Pamela J Benoit. 1992. Perspectives on argument. Readings in argumentation, pages 121–143. find
  95. Sarah Wiegreffe, Ana Marasović, and Noah A Smith. 2020. Measuring association between labels and free-text rationales. arXiv preprint arXiv:2010.12762.
  96. Yongjin Yang, Joonkee Kim, Yujin Kim, Namgyu Ho, James Thorne, and Se-Young Yun. 2023. HARE: Explainable hate speech detection with step-by-step reasoning. arXiv [cs.CL]. find
  97. Xi Ye and Greg Durrett. 2022. The unreliability of explanations in few-shot prompting for textual reasoning. Advances in neural information processing systems, 35:30378–30392.
  98. Zhangyue Yin, Qiushi Sun, Qipeng Guo, Zhiyuan Zeng, Xiaonan Li, Junqi Dai, Qinyuan Cheng, Xuan-Jing Huang, and Xipeng Qiu. 2024. Reasoning in flux: Enhancing large language models reasoning through uncertainty-aware adaptive guidance. In Proceedings of the 62nd Annual Meeting of the Association for… find
  99. Zhehao Zhang, Jiaao Chen, and Diyi Yang. 2023. Mitigating Biases in Hate Speech Detection from A Causal Perspective. In Findings of the Association for Computational Linguistics: EMNLP 2023, pages 6610–6625, Singapore. Association for Computational Linguistics. find
  100. Haiyan Zhao, Hanjie Chen, Fan Yang, Ninghao Liu, Huiqi Deng, Hengyi Cai, Shuaiqiang Wang, Dawei Yin, and Mengnan Du. 2024a. Explainability for large language models: A survey. ACM Transactions on Intelligent Systems and Technology, 15(2):1–38.
  101. Lingjun Zhao and Hal Daumé III. 2025. A necessary step toward faithfulness: Measuring and improving consistency in free-text explanations. arXiv preprint arXiv:2505.19299. find
  102. Yibo Zhao, Jiapeng Zhu, Can Xu, and Xiang Li. 2024b. Enhancing llm-based hatred and toxicity detection with meta-toxic knowledge graph. arXiv preprint arXiv:2412.15268. find
  103. Tianshi Zheng, Yixiang Chen, Chengxi Li, Chunyang Li, Qing Zong, Haochen Shi, Baixuan Xu, Yangqiu Song, Ginny Y Wong, and Simon See. 2025. The curse of cot: On the limitations of chain-of-thought in in-context learning. arXiv preprint arXiv:2504.05081. — [arXiv:2504.05081](https://arxiv.org/abs/2504.05081) find
  104. Jin Peng Zhou, Charles Staats, Wenda Li, Christian Szegedy, Kilian Q Weinberger, and Yuhuai Wu. 2024. Don’t trust: Verify–grounding llm quantitative reasoning with autoformalization. arXiv preprint arXiv:2403.18120. find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/mothilal-2026-argument-based-consistency-toxicity-explanations/