¿Inferencias en máquinas? Alucinaciones inferenciales de los grandes modelos de lenguaje desde una perspectiva de los sentimientos metacognitivos

Palabras clave: alucinaciones inferenciales, grandes modelos de lenguaje, sentimientos metacognitivos

Resumen

Los grandes modelos de lenguaje han ganado recientemente cierta popularidad debido a su habilidad para hacer inferencias a partir de la información proporcionada por los usuarios. Estas redes neuronales parecen inferir lógicamente, sobre una base probabilística, emulando reglas lógicas. Tradicionalmente, las inferencias han sido entendidas como procesos de pensamiento que van desde las premisas hasta las conclusiones. No obstante, se ha debatido sobre la naturaleza de la inferencia y la posibilidad de extenderla por fuera del ámbito humano: ¿es comparable la capacidad inferencial de los grandes modelos de lenguaje con la de los seres humanos? En este artículo se propone que la diferencia entre la inferencia humana y la de los grandes modelos de lenguaje radica en la dimensión fenomenológica o afectiva del razonamiento. En particular, la ausencia de sentimientos metacognitivos que guían, monitorean y evalúan los procesos de pensamiento; por ejemplo, razonar o inferir. Aunque los grandes modelos de lenguaje simulan inferencias novedosas, su tendencia a producir alucinaciones inferenciales demuestra la ausencia de esa dimensión autorreguladora que caracteriza al razonamiento humano. En contraste con la afectividad presente en los humanos, las fallas inferenciales de los grandes modelos de lenguaje se podrían entender a partir de tres aspectos: la maldición de la reversión, las alucinaciones inferenciales por estrategias de decodificación y el exceso de confianza. Este artículo concluye que unas de las causas de las alucinaciones inferenciales son la falta de sentimientos metacognitivos como el sentimiento de corrección, la falta de balance por sentimiento de fiabilidad y la ausencia de la duda.

Biografía del autor/a

Sebastián Mejía-Rendón, Universidad Nacional de Córdoba

Córdoba, Argentina, joan.mejia@mi.unc.edu.ar

Mariana Olezza, Universidad de Buenos Aires

Buenos Aires, Argentina, olezza.mur@gmail.com

Referencias bibliográficas

Arango-Muñoz, S. (2011). Two Levels of Metacognition. Philosophia, 39(1), 71-82. https://doi.org/10.1007/s11406-010-9279-0

Arango-Muñoz, S., & Michaelian, K. (2014). Epistemic Feelings, Epistemic Emotions: Review and Introduction to the Focus Section. Philosophical Inquiries, 2(1), 97-122. https://doi.org/10.4454/philinq.v2i1.79

Basmov, V., Goldberg, Y., & Tsarfaty, R. (2023). Simple Linguistic Inferences of Large Language Models (LLMs): Blind Spots and Blinds. arXiv. https://doi.org/10.48550/arXiv.2305.14785

Bellucci, F., & Pietarinen, A. V. (2023). Peirce’s Abduction. En L. Magnani (Ed.), Handbook of Abductive Cognition (pp. 7-20). Springer. https://doi.org/10.1007/978-3-031-10135-9_7

Bender, E. M., Gebru, T., McMillan-Major, A., & Shmitchell, S. (2021). On the Dangers of Stochastic Parrots. En Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency (pp. 610-623). Association for Computing Machinery. https://doi.org/10.1145/3442188.3445922

Berglund, L., Tong, M., Kaufmann, M., Balesni, M., Stickland, A. C., Korbak, T., & Evans, O. (2023). The Reversal Curse: LLMs trained on “A is B” fail to learn “B is A”. arXiv. https://doi.org/10.48550/arXiv.2309.12288

Boghossian, P. (2014). What is inference? Philosophical Studies, 169(1), 1-18. https://doi.org/10.1007/s11098-012-9903-x

Boghossian, P. (2018). Delimiting the Boundaries of Inference. Philosophical Issues, 28(1), 55-69. https://doi.org/10.1111/phis.12115

Borji, A. (2023). A Categorical Archive of ChatGPT Failures. arXiv. https://doi.org/10.48550/arXiv.2302.03494

Boyle, M. (2011). Transparent Self-Knowledge. Aristotelian Society, 85(1), 223-241. https://doi.org/10.1111/j.1467-8349.2011.00204.x

Broome, J. (2013). Practical Reasoning and Inference. En D. Bakhurst, B. Hooker & M. O. Little (Eds.), Thinking About Reasons: Themes from the Philosophy of Jonathan Dancy (pp. 286-309). Oxford University Press. https://doi.org/10.1093/acprof:oso/9780199604678.003.0014

Broome, J. (2014). Normativity in Reasoning. Pacific Philosophical Quarterly, 95(4), 622-633. https://doi.org/10.1111/papq.12050

Brown, A. S. (1991). A review of the tip-of-the-tongue experience. Psychological Bulletin, 109(2), 204-223. https://doi.org/10.1037/0033-2909.109.2.204

Bryson, A. E., Ho, Y.-C., & Siouris, M. (1969). Applied Optimal Control: Optimization, Estimation, and Control. IEEE Transactions on Systems, Man, and Cybernetics, 9(6), 366-367. https://doi.org/10.1109/TSMC.1979.4310229

Burda, Y., Edwards, H., Pathak, D., Storkey, A., Darrell, T., & Efros, A. A. (2018). Large-Scale Study of Curiosity-Driven Learning. arXiv. https://doi.org/10.48550/arXiv.1808.04355

Chen, X., Li, M., Gao, X., & Zhang, X. (2022). Towards improving faithfulness in abstractive summarization. En Proceedings of the 36th International Conference on Neural Information Processing Systems (NIPS '22) (pp. 24516-24528, art. 1780). Curran Associates Inc. https://dl.acm.org/doi/10.5555/3600270.3602050

Chen, Y., Singh, V. K., Ma, J., & Tang, R. (2025). CounterBench: A Benchmark for Counterfactuals Reasoning in Large Language Models. arXiv. https://doi.org/10.48550/arXiv.2502.11008

Chudnoff, E. (2014). Intuition in Mathematics. En L. M. Osbeck & B. S. Held (Eds.), Rational Intuition: Philosophical Roots, Scientific Investigations (pp. 174-191). Cambridge University Press. https://doi.org/10.1017/CBO9781139136419.010

Chudnoff, E. (2024). Inferential Seemings. En U. Kriegel (Ed.), Oxford Studies in Philosophy of Mind: Volume 4 (pp. 311-333). Oxford. https://doi.org/10.1093/9780198924159.003.0009

de Sousa, R. (2009). Epistemic Feelings. Mind & Matter, 7(2), 139-161. https://www.researchgate.net/publication/233488784_Epistemic_Feelings

Dokic, J. (2012). Seeds of self-knowledge: Noetic feelings and metacognition. En M. J. Beran, J. L. Brandl, J. Perner & J. Proust (Eds.), Foundations of Metacognition (pp. 302-321). Oxford University Press. https://doi.org/10.1093/acprof:oso/9780199646739.003.0020

Dreyfus, H. L. (1972). What Computers Can’t Do: A Critique of Artificial Reason. Harper & Row.

Eschmann, J. (2021). Reward Function Design in Reinforcement Learning. En B. Belousov, H. Abdulsamad, P. Klink, S. Parisi & J. Peters (Eds.), Reinforcement Learning Algorithms: Analysis and Applications (pp. 25-33). Springer.

Fernandez Cruz, A. L., Arango-Muñoz, S., & Volz, K. G. (2016). Oops, scratch that! Monitoring one’s own errors during mental calculation. Cognition, 146, 110-120. https://doi.org/10.1016/j.cognition.2015.09.005

Fernández Velasco, P. (2022). Disorientation and self-consciousness: a phenomenological inquiry. Phenomenology and the Cognitive Sciences, 21(1), 203-222. https://doi.org/10.1007/s11097-020-09659-1

Fernandez Velasco, P., & Loev, S. (2022). Cognitive Feelings in the Predictive Mind: Emotion, Meta-cognition and Predictive Processing. PsyArXiv. https://doi.org/10.31234/osf.io/594hj

Gallagher, S. (2005). How the Body Shapes the Mind. Oxford University Press. https://doi.org/10.1093/0199271941.001.0001

Goodfellow, I., Bengio, Y., & Courville, A. (2016). Deep Learning. The MIT Press.

Goupil, L., & Proust, J. (2023). Curiosity as a metacognitive feeling. Cognition, 231, art. 105325. https://doi.org/10.1016/j.cognition.2022.105325

Hart, J. T. (1965). Memory and the feeling-of-knowing experience. Journal of Educational Psychology, 56(4), 208-216. https://doi.org/10.1037/h0022263

Hill-Yardin, E. L., Hutchinson, M. R., Laycock, R., & Spencer, S. J. (2023). A Chat (GPT) about the future of scientific publishing. Brain, Behavior and Immunity, 110, 152-154. https://doi.org/10.1016/j.bbi.2023.02.022

Hlobil, U. (2014). Against Boghossian, Wright and Broome on inference. Philosophical Studies, 167(2), 419-429. https://doi.org/10.1007/s11098-013-0104-z

Huang, L., Yu, W., Ma, W., Zhong, W., Feng, Z., Wang, H., Chen, Q., Peng, W., Feng, X., Qin, B., & Liu, T. (2023). A Survey on Hallucination in Large Language Models: Principles, Taxonomy, Challenges, and Open Questions. ACM Transactions on Information Systems, 43(2), art. 42. https://doi.org/10.1145/3703155

Kalouli, A. L., Buis, A., Real, L., Palmer, M., & De Paiva, V. (2019). Explaining simple natural language inference. En Proceedings of the 13th Linguistic Annotation Workshop (pp. 132-143). Association for Computational Linguistics.

Kornblith, H. (2012). On Reflection. Oxford University Press.

Larson, E. J. (2021). The Myth of Artificial Intelligence: Why Computers Can’t Think the Way We Do. Harvard University Press.

Lin, C.-Y. (2004). ROUGE: A Package for Automatic Evaluation of Summaries. En Text Summarization Branches Out (pp. 74-81). Association for Computational Linguistics. https://aclanthology.org/W04-1013.pdf

Lin, S., Hilton, J., & Evans, O. (2022). TruthfulQA: Measuring How Models Mimic Human Falsehoods. En S. Muresan, P. Nakov & A. Villavicencio (Eds.), Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers) (pp. 3214-3252). Association for Computational Linguistics. https://doi.org/10.18653/v1/2022.acl-long.229

Manzotti, R., & Chella, A. (2018). Good Old-Fashioned Artificial Consciousness and the Intermediate Level Fallacy. Frontiers in Robotics and AI, 5, art. 39. https://doi.org/10.3389/frobt.2018.00039

Markovits, H., & Vachon, R. (1989). Reasoning with contrary-to-fact propositions. Journal of Experimental Child Psychology, 47(3), 398-412. https://doi.org/10.1016/0022-0965(89)90021-0

Mejía-Rendón, J. S., & Olezza, M. (2024). ChatGPT e Inferencias. En S. Stein & V. Braga (Eds.), Libro de resúmenes: XIII Encuentro de filosofía e historia de la ciencia del Cono Sur 2024 (pp. 170-171). Editorial Philosophia. http://www.afhic.com/wp-content/uploads/2024/06/Caderno-de-Resumos-AFHIC-2024-Vers%C3%A3o-final.pdf

Metcalfe, J., & Shimamura, A. P. (Eds.). (1994). Metacognition: Knowing about Knowing. The MIT Press. https://doi.org/10.7551/mitpress/4561.001.0001

Moshman, D., & Franks, B. A. (1986). Development of the Concept of Inferential Validity. Child Development, 57(1), 153-165. https://doi.org/10.2307/1130647

Mukherjee, A. (2024). Psittacines of Innovation? Assessing the True Novelty of AI Creations. arXiv. https://doi.org/10.48550/arXiv.2404.00017

Olezza, M. (2025). Analogical mappings of facts and counterfactuals in the human mind and Peirce’s abduction: limitations in LLMs. Cognitive Systems Research, 94, art. 101408. https://doi.org/10.1016/j.cogsys.2025.101408

Pearl, J. (2019). The seven tools of causal inference, with reflections on machine learning. Communications of the ACM, 62(3), 54-60. https://doi.org/10.1145/3241036

Pearl, J., & Mackenzie, M. (2018). The Book of Why. Basic Books.

Proust, J. (2013). The Philosophy of Metacognition: Mental Agency and Self-Awareness. Oxford University Press. https://doi.org/10.1093/acprof:oso/9780199602162.001.0001

Proust, J., & Pliushch, I. (2016). The Representational Structure of Feelings. En T. Metzinger & J. M. Windt (Eds.), Open MIND, 2-vol. set: Philosophy and the Mind Sciences in the 21st Century. The MIT Press. https://doi.org/10.7551/mitpress/10603.003.0094

Quilty-Dunn, J., & Mandelbaum, E. (2018). Against dispositionalism: belief in cognitive science. Philosophical Studies, 175(9), 2353-2372. https://doi.org/10.1007/s11098-017-0962-x

Ruder, S. (2016). An overview of gradient descent optimization algorithms. arXiv. https://doi.org/10.48550/arXiv.1609.04747

Rumelhart, D. E., & McClelland, J. L. (1986). Parallel Distributed Processing: Explorations in the Microstructure of Cognition. The MIT Press.

Ryan, R. M., & Deci, E. L. (2000). Intrinsic and Extrinsic Motivations: Classic Definitions and New Directions. Contemporary Educational Psychology, 25(1), 54-67. https://doi.org/10.1006/ceps.1999.1020

Searle, J. R. (1980). Minds, Brains and Programs. Behavioral and Brain Sciences, 3(3), 417-457. https://doi.org/10.1017/S0140525X00005756

Shea, N. (2024). Metacognition of Inferential Transitions. Journal of Philosophy, 121(11), 597-627. https://doi.org/10.5840/jphil20246202

Stetie, N. A. (2023). ¿Con qué lentes buscamos a Wally? Articulaciones y problemas de interfaz entre la lingüística y el estudio de la cognición [conferencia]. Workshop de Estudios sobre Ciencias Cognitivas y Neurociencias, Córdoba, Argentina. https://www.conicet.gov.ar/new_scp/detalle.php?keywords=&id=56674&congresos=yes&detalles=yes&congr_id=11015032

Strawson, G. (2003). Mental Ballistics or the Involuntariness of Spontaneity. Proceedings of the Aristotelian Society, 103(1), 227-256. https://doi.org/10.1111/j.0066-7372.2003.00071.x

Thompson, V. A., Prowse Turner, J. A., & Pennycook, G. (2011). Intuition, reason, and metacognition. Cognitive Psychology, 63(3), 107-140. https://doi.org/10.1016/j.cogpsych.2011.06.001

Valaris, M. (2017). What reasoning might be. Synthese, 194(6), 2007-2024. https://doi.org/10.1007/s11229-016-1034-z

Vaswani, A., Shazeer, N., Parmar, N., Uszkoreit, J., Jones, L., Gomez, A. N., Kaiser, L., & Polosukhin, I. (2017). Attention Is All You Need. En Proceedings of the 31st International Conference on Neural Information Processing Systems (NIPS'17) (pp. 6000-6010). Curran Associates Inc. https://dl.acm.org/doi/10.5555/3295222.3295349

Wright, C. (2014). Comment on Paul Boghossian, “What is inference”. Philosophical Studies, 169(1), 27-37. https://doi.org/10.1007/s11098-012-9892-9

Yang, L., Li, J., Cunningham, P., Zhang, Y., Smyth, B., & Dong, R. (2021). Exploring the Efficacy of Automatically Generated Counterfactuals for Sentiment Analysis. En C. Zong, F. Xia, W. Li & R. Navigli (Eds.), Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers) (pp. 306-316). Association for Computational Linguistics. https://doi.org/10.18653/v1/2021.acl-long.26

Zahavi, D. (2025). Phenomenology: The Basics (2.a ed.). Routledge.

Cómo citar
Mejía-Rendón, S., & Olezza, M. (2026). ¿Inferencias en máquinas? Alucinaciones inferenciales de los grandes modelos de lenguaje desde una perspectiva de los sentimientos metacognitivos. Trilogía Ciencia Tecnología Sociedad, 18(38), e3702. https://doi.org/10.22430/21457778.3702

Descargas

Los datos de descargas todavía no están disponibles.
Publicado
08-04-2026
Crossref Cited-by logo
Escanea para compartir
Código QR para abrir y compartir esta publicación
Estadísticas de artículo
Vistas de resúmenes 470
Vistas de PDF 229
Descargas de PDF 178
Vistas de HTML 20
Otras vistas 31