mlsanityData to 7 Oct 2026

Confabulation: The Surprising Value of Large Language Model Hallucinations

arXiv 2406.04175 ↗First posted 6 Jun 2024cs.CL cs.AIcited by 42

Not on the map. Why?

Abstract (arXiv)

This paper presents a systematic defense of large language model (LLM) hallucinations or 'confabulations' as a potential resource instead of a categorically negative pitfall. The standard view is that confabulations are inherently problematic and AI research should eliminate this flaw. In this paper, we argue and empirically demonstrate that measurable semantic characteristics of LLM confabulations mirror a human propensity to utilize increased narrativity as a cognitive resource for sense-making and communication. In other words, it has potential value. Specifically, we analyze popular hallucination benchmarks and reveal that hallucinated outputs display increased levels of narrativity and semantic coherence relative to veridical outputs. This finding reveals a tension in our usually dismissive understandings of confabulation. It suggests, counter-intuitively, that the tendency for LLMs to confabulate may be intimately associated with a positive capacity for coherent narrative-text generation.

Key facts. Cited by 42 papers in the map (data through 7 Oct 2026). Builds on 27 of its 81 references that are in the map. First posted on arXiv 6 Jun 2024.

Builds on

27 of 81 references.

References · 27

  1. 2022Training language models to follow instructions with human feedbackcited by 15,712
  2. 2023Llama 2: Open Foundation and Fine-Tuned Chat Modelscited by 12,865
  3. 2023A Survey on Hallucination in Large Language Models: Principles, Taxonomy, Challenges, and Open Questionscited by 1,582
  4. 2018Wizard of Wikipedia: Knowledge-Powered Conversational agentscited by 530
  5. 2023LLM Lies: Hallucinations are not Bugs, but Features as Adversarial Examplescited by 177
  6. 2023Med-HALT: Medical Domain Hallucination Test for Large Language Modelscited by 156
  7. 2022On the Origin of Hallucinations in Conversational Models: Is it the Datasets or the Models?cited by 137
  8. 2021Hallucinated but Factual! Inspecting the Factuality of Hallucinations in Abstractive Summarizationcited by 128
  9. 2022FaithDial: A Faithful Benchmark for Information-Seeking Dialoguecited by 108
  10. 2023Cognitive Dissonance: Why Do Language Model Outputs Disagree with Internal Representations of Truthfulness?cited by 62
  11. 2023Deficiency of Large Language Models in Finance: An Empirical Examination of Hallucinationcited by 41
  12. 2023Scene Graph as Pivoting: Inference-time Image-free Unsupervised Multimodal Machine Translation with Visual Scene Hallucinationcited by 27
15 more
  1. 2022DEAM: Dialogue Coherence Evaluation using AMR-based Semantic Manipulationscited by 25
  2. 2023HaLP: Hallucinating Latent Positives for Skeleton-based Self-Supervised Learning of Actionscited by 25
  3. 2023The Dark Side of ChatGPT: Legal and Ethical Challenges from Stochastic Parrots and Hallucinationcited by 23
  4. 2024Hallucination Detection and Hallucination Mitigation: An Investigationcited by 20
  5. 2023Hallucination Improves the Performance of Unsupervised Visual Representation Learningcited by 16
  6. 2023Where Do People Tell Stories Online? Story Detection Across Online Communitiescited by 16
  7. 2023Hallucination is the last thing you needcited by 11
  8. 2022Deep Learning Opacity in Scientific Discoverycited by 9
  9. 2023From human-centered to social-centered artificial intelligence: Assessing ChatGPT's impact through disruptive eventscited by 8
  10. 2024Predicting challenge moments from students' discourse: A comparison of GPT-4 to two traditional natural language processing approachescited by 8
  11. 2023User-Controlled Knowledge Fusion in Large Language Models: Balancing Creativity and Hallucinationcited by 7
  12. 2021Multi-Object Tracking with Hallucinated and Unlabeled Videoscited by 6
  13. 2023The Creative Frontier of Generative AI: Managing the Novelty-Usefulness Tradeoffcited by 6
  14. 2024Oil & Water? Diffusion of AI Within and Across Scientific Fieldscited by 6
  15. 2023Correction with Backtracking Reduces Hallucination in Summarizationcited by 4

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2024: 6; 2025: 22; 2026: 14.

42 papers.

Sort:

By area:

  1. 2025ELEPHANT: Measuring and understanding social sycophancy in LLMscited by 115
  2. 2024StructRAG: Boosting Knowledge Intensive Reasoning of LLMs via Inference-time Hybrid Information Structurizationcited by 70
  3. 2024DocETL: Agentic Query Rewriting and Evaluation for Complex Document Processingcited by 49
  4. 2024Large Language Models and Cognitive Science: A Comprehensive Review of Similarities, Differences, and Challengescited by 48
  5. 2025Technological folie `a deux: Feedback Loops Between AI Chatbots and Mental Illnesscited by 14
  6. 2025Thinking beyond the anthropomorphic paradigm benefits LLM researchcited by 13
  7. 2024On Classification with Large Language Models in Cultural Analyticscited by 9
  8. 2026CreativeBench: Benchmarking and Enhancing Machine Creativity via Self-Evolving Challengescited by 9
  9. 2025Provocations from the Humanities for Generative AI Researchcited by 8
  10. 2026Accommodation and Epistemic Vigilance: A Pragmatic Account of Why LLMs Fail to Challenge Harmful Beliefscited by 8
  11. 2025Why Slop Matterscited by 6
  12. 2025KRISTEVA: Close Reading as a Novel Task for Benchmarking Interpretive Reasoningcited by 4
30 more
  1. 2025Can LLMs Detect Their Confabulations? Estimating Reliability in Uncertainty-Aware Language Modelscited by 4
  2. 2025ReFACT: A Benchmark for Scientific Confabulation Detection with Positional Error Annotationscited by 4
  3. 2025PENDULUM: A Benchmark for Assessing Sycophancy in Multimodal Large Language Modelscited by 4
  4. 2024Towards a Science Exocortexcited by 3
  5. 2026Semantic Calibration Prevails Where Token Confidence Fails: Benchmarking Long-Form Scientific QAcited by 3
  6. 2026Narrative Theory-Driven LLM Methods for Automatic Story Generation and Understanding: A Surveycited by 3
  7. 2026The Dynamics of Delusion: Modeling Bidirectional False Belief Amplification in Human-Chatbot Dialoguecited by 3
  8. 2025Memory Helps, but Confabulation Misleads: Understanding Streaming Events in Videos with MLLMscited by 2
  9. 2025Synthetic Fluency: Hallucinations, Confabulations, and the Creation of Irish Words in LLM-Generated Translationscited by 2
  10. 2025MetaRAG: Metamorphic Testing for Hallucination Detection in RAG Systemscited by 2
  11. 2025Explainability of Large Language Models: Opportunities and Challenges toward Generating Trustworthy Explanationscited by 2
  12. 2025Critical Confabulation: Can LLMs Hallucinate for Social Good?cited by 2
  13. 2026Structure Guided Retrieval-Augmented Generation for Factual Queriescited by 2
  14. 2024AiBAT: Artificial Intelligence/Instructions for Build, Assembly, and Testcited by 1
  15. 2025OAEI-LLM-T: A TBox Benchmark Dataset for Understanding Large Language Model Hallucinations in Ontology Matchingcited by 1
  16. 2025Personalized Constitutionally-Aligned Agentic Superego: Secure AI Behavior Aligned to Diverse Human Valuescited by 1
  17. 2025Library Hallucinations in LLM-Generated Code: A Risk Analysis Grounded in Developer Queriescited by 1
  18. 2026The Homogenization Problem in LLMs: Towards Meaningful Diversity in AI Safetycited by 1
  19. 2025ArxEval: Evaluating Retrieval and Generation in Language Models for Scientific Literaturecited by 0
  20. 2025Valuable Hallucinations: Realizable Non-realistic Propositionscited by 0
  21. 2025`Generalization is hallucination' through the lens of tensor completionscited by 0
  22. 2025Neither Stochastic Parroting nor AGI: LLMs Solve Tasks through Context-Directed Extrapolation from Training Data Priorscited by 0
  23. 2025LongRecall: A Structured Approach for Robust Recall Evaluation in Long-Form Textcited by 0
  24. 2026Fabricator or dynamic translator?cited by 0
  25. 2026Not All That Is Fluent Is Factual: Investigating Hallucinations of Large Language Models in Academic Writingcited by 0
  26. 2026IDEAFix: Evaluation Framework for Creative Defixation Prompting in LLMscited by 0
  27. 2026AI Virtue: What is "Good" Knowledge in the Age of Artificial Intelligence?cited by 0
  28. 2026Optimization Is Not All You Needcited by 0
  29. 2026MemArena: An Ego-Centric Benchmark for On-Device Agentic Personal Memory Assistants at Scalecited by 0
  30. 2026Auditing the Synthetic Memoir: Measuring Scene-Level Confabulation in LLM-Generated Autobiography Against the Documented Record of the Life It Describescited by 0