mlsanityData to 14 Oct 2026

Review of Hallucination Understanding in Large Language and Vision Models

arXiv 2510.00034 ↗First posted 26 Sep 2025cs.CV cs.AIcited by 2

Not on the map. Why?

Abstract (arXiv)

The widespread adoption of large language and vision models in real-world applications has made urgent the need to address hallucinations -- instances where models produce incorrect or nonsensical outputs. These errors can propagate misinformation during deployment, leading to both financial and operational harm. Although much research has been devoted to mitigating hallucinations, our understanding of it is still incomplete and fragmented. Without a coherent understanding of hallucinations, proposed solutions risk mitigating surface symptoms rather than underlying causes, limiting their effectiveness and generalizability in deployment. To tackle this gap, we first present a unified, multi-level framework for characterizing both image and text hallucinations across diverse applications, aiming to reduce conceptual fragmentation. We then link these hallucinations to specific mechanisms within a model's lifecycle, using a task-modality interleaved approach to promote a more integrated understanding. Our investigations reveal that hallucinations often stem from predictable patterns in data distributions and inherited biases. By deepening our understanding, this survey provides a foundation for developing more robust and effective solutions to hallucinations in real-world generative AI systems.

Key facts. Cited by 2 papers in the map (data through 14 Oct 2026). Builds on 138 of its 187 references that are in the map. First posted on arXiv 26 Sep 2025.

Builds on

138 of 187 references.

References · 138

  1. 2023Judging LLM-as-a-Judge with MT-Bench and Chatbot Arenacited by 8,626
  2. 2023Direct Preference Optimization: Your Language Model is Secretly a Reward Modelcited by 8,065
  3. 2023Lost in the Middle: How Language Models Use Long Contextscited by 3,455
  4. 2021TruthfulQA: Measuring How Models Mimic Human Falsehoodscited by 2,975
  5. 2023Efficient Streaming Language Models with Attention Sinkscited by 2,186
  6. 2020Shortcut Learning in Deep Neural Networkscited by 1,969
  7. 2020Transformer Feed-Forward Layers Are Key-Value Memoriescited by 1,336
  8. 2023LIMA: Less Is More for Alignmentcited by 1,100
  9. 2023Large Language Models are not Fair Evaluatorscited by 1,085
  10. 2019Aleatoric and Epistemic Uncertainty in Machine Learning: An Introduction to Concepts and Methodscited by 913
  11. 2023A Multitask, Multilingual, Multimodal Evaluation of ChatGPT on Reasoning, Hallucination, and Interactivitycited by 894
  12. 2023Extracting Training Data from Diffusion Modelscited by 732
38 more
  1. 2024LLM Evaluators Recognize and Favor Their Own Generationscited by 686
  2. 2023An Empirical Study of Catastrophic Forgetting in Large Language Models During Continual Fine-tuningcited by 639
  3. 2022When and why vision-language models behave like bags-of-words, and what to do about it?cited by 610
  4. 2023Dissecting Recall of Factual Associations in Auto-Regressive Language Modelscited by 528
  5. 2023Are Emergent Abilities of Large Language Models a Mirage?cited by 523
  6. 2023Faith and Fate: Limits of Transformers on Compositionalitycited by 503
  7. 2022Large Language Models Struggle to Learn Long-Tail Knowledgecited by 458
  8. 2022Diffusion Art or Digital Forgery? Investigating Data Replication in Diffusion Modelscited by 424
  9. 2023The Reversal Curse: LLMs trained on "A is B" fail to learn "B is A"cited by 407
  10. 2024Justice or Prejudice? Quantifying Biases in LLM-as-a-Judgecited by 368
  11. 2023A Comprehensive Survey on Test-Time Adaptation under Distribution Shiftscited by 359
  12. 2020The Pitfalls of Simplicity Bias in Neural Networkscited by 351
  13. 2023The Impact of Positional Encoding on Length Generalization in Transformerscited by 327
  14. 2023Studying Large Language Model Generalization with Influence Functionscited by 326
  15. 2024Humans or LLMs as the Judge? A Study on Judgement Biasescited by 323
  16. 2023Rethinking FID: Towards a Better Evaluation Metric for Image Generationcited by 310
  17. 2023Self-Consuming Generative Models Go MADcited by 250
  18. 2023Reasoning or Reciting? Exploring the Capabilities and Limitations of Language Models Through Counterfactual Taskscited by 246
  19. 2023Proving Test Set Contamination in Black Box Language Modelscited by 230
  20. 2024Leak, Cheat, Repeat: Data Contamination and Evaluation Malpractices in Closed-Source LLMscited by 229
  21. 2023Understanding and Mitigating Copying in Diffusion Modelscited by 209
  22. 2024Transcoders Find Interpretable LLM Feature Circuitscited by 200
  23. 2023Benchmarking Foundation Models with Language-Model-as-an-Examinercited by 172
  24. 2023Quantizable Transformers: Removing Outliers by Helping Attention Heads Do Nothingcited by 151
  25. 2023Benchmarking Cognitive Biases in Large Language Models as Evaluatorscited by 150
  26. 2023Emergent and Predictable Memorization in Large Language Modelscited by 148
  27. 2024The PRISM Alignment Dataset: What Participatory, Representative and Individualised Human Feedback Reveals About the Subjective and Multicultural Alignment of Large Language Modelscited by 144
  28. 2024Scaling Laws for Reward Model Overoptimization in Direct Alignment Algorithmscited by 143
  29. 2024ODIN: Disentangled Reward Mitigates Hacking in RLHFcited by 136
  30. 2024A Tale of Tails: Model Collapse as a Change of Scaling Lawscited by 130
  31. 2024Cooperate or Collapse: Emergence of Sustainable Cooperation in a Society of LLM Agentscited by 125
  32. 2024MaxMin-RLHF: Alignment with Diverse Human Preferencescited by 123
  33. 2025Red-Teaming LLM Multi-Agent Systems via Communication Attackscited by 122
  34. 2024Secret Collusion among AI Agents: Multi-Agent Deception via Steganographycited by 113
  35. 2022Why Exposure Bias Matters: An Imitation Learning Perspective of Error Accumulation in Language Generationcited by 108
  36. 2023Elucidating the Exposure Bias in Diffusion Modelscited by 101
  37. 2024Found in the Middle: Calibrating Positional Attention Bias Improves Long Context Utilizationcited by 101
  38. 2023On the Stability of Iterative Retraining of Generative Models on their own Datacited by 100

50 most cited of 138 All 138 (JSON)

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 2.

2 papers.

  1. 2026One-shot Optimized Steering Vector for Hallucination Mitigation for VLMscited by 1
  2. 2026Spilled Energy in Large Language Modelscited by 1