mlsanityData to 7 Oct 2026

Unsupervised Real-Time Hallucination Detection based on the Internal States of Large Language Models

arXiv 2403.06448 ↗First posted 11 Mar 2024cs.CL cs.AIcited by 84

Not on the map. Why?

Abstract (arXiv)

Hallucinations in large language models (LLMs) refer to the phenomenon of LLMs producing responses that are coherent yet factually inaccurate. This issue undermines the effectiveness of LLMs in practical applications, necessitating research into detecting and mitigating hallucinations of LLMs. Previous studies have mainly concentrated on post-processing techniques for hallucination detection, which tend to be computationally intensive and limited in effectiveness due to their separation from the LLM's inference process. To overcome these limitations, we introduce MIND, an unsupervised training framework that leverages the internal states of LLMs for real-time hallucination detection without requiring manual annotations. Additionally, we present HELM, a new benchmark for evaluating hallucination detection across multiple LLMs, featuring diverse LLM outputs and the internal states of LLMs during their inference process. Our experiments demonstrate that MIND outperforms existing state-of-the-art methods in hallucination detection.

Key facts. Cited by 84 papers in the map (data through 7 Oct 2026). Builds on 30 of its 52 references that are in the map. First posted on arXiv 11 Mar 2024.

Builds on

30 of 52 references.

References · 30

  1. 2020Language Models are Few-Shot Learnerscited by 32,199
  2. 2023LLaMA: Open and Efficient Foundation Language Modelscited by 13,958
  3. 2020Retrieval-Augmented Generation for Knowledge-Intensive NLP Taskscited by 9,160
  4. 2022PaLM: Scaling Language Modeling with Pathwayscited by 5,304
  5. 2022OPT: Open Pre-trained Transformer Language Modelscited by 3,464
  6. 2016Pointer Sentinel Mixture Modelscited by 2,992
  7. 2021TruthfulQA: Measuring How Models Mimic Human Falsehoodscited by 2,919
  8. 2020The Pile: An 800GB Dataset of Diverse Text for Language Modelingcited by 2,346
  9. 2020REALM: Retrieval-Augmented Language Model Pre-Trainingcited by 2,126
  10. 2022BLOOM: A 176B-Parameter Open-Access Multilingual Language Modelcited by 1,968
  11. 2022Language Models (Mostly) Know What They Knowcited by 1,783
  12. 2020Leveraging Passage Retrieval with Generative Models for Open Domain Question Answeringcited by 1,203
18 more
  1. 2019Generalization through Memorization: Nearest Neighbor Language Modelscited by 788
  2. 2023REPLUG: Retrieval-Augmented Black-Box Language Modelscited by 786
  3. 2023The RefinedWeb Dataset for Falcon LLM: Outperforming Curated Corpora with Web Data, and Web Data Onlycited by 755
  4. 2023The Internal State of an LLM Knows When It's Lyingcited by 644
  5. 2020Detecting Hallucinated Content in Conditional Neural Sequence Generationcited by 151
  6. 2021A Token-level Reference-free Hallucination Detection Benchmark for Free-form Text Generationcited by 137
  7. 2023Enhancing Uncertainty-Based Hallucination Detection with Stronger Focuscited by 98
  8. 2022Retrieval as Attention: End-to-end Learning of Retrieval and Reading within a Single Transformercited by 45
  9. 2023Towards Better Web Search Performance: Pre-training, Fine-tuning and Learning to Rankcited by 17
  10. 2023THUIR@COLIEE 2023: Incorporating Structural Knowledge into Pre-trained Language Models for Legal Case Retrievalcited by 16
  11. 2023From Relevance to Utility: Evidence Retrieval with Feedback for Fact Verificationcited by 15
  12. 2023Wikiformer: Pre-training with Structured Information of Wikipedia for Ad-hoc Retrievalcited by 15
  13. 2023CaseEncoder: A Knowledge-enhanced Pre-trained Model for Legal Case Encodingcited by 14
  14. 2023THUIR2 at NTCIR-16 Session Search (SS) Taskcited by 12
  15. 2023THUIR at WSDM Cup 2023 Task 1: Unbiased Learning to Rankcited by 11
  16. 2023THUIR@COLIEE 2023: More Parameters and Legal Knowledge for Legal Case Entailmentcited by 11
  17. 2023Caseformer: Pre-training for Legal Case Retrieval Based on Inter-Case Distinctionscited by 9
  18. 2019Healthcare NER Models Using Language Model Pretrainingcited by 8

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2024: 17; 2025: 34; 2026: 33.

84 papers.

Sort:

By area:

  1. 2024Devils in Middle Layers of Large Vision-Language Models: Interpreting, Detecting and Mitigating Object Hallucinations via Attention Lenscited by 139
  2. 2024Interpreting and Editing Vision-Language Representations to Mitigate Hallucinationscited by 103
  3. 2024HaloScope: Harnessing Unlabeled LLM Generations for Hallucination Detectioncited by 96
  4. 2024Trustworthiness in Retrieval-Augmented Generation Systems: A Surveycited by 83
  5. 2024Latent Space Chain-of-Embedding Enables Output-free LLM Self-Evaluationcited by 56
  6. 2025HaluMem: Evaluating Hallucinations in Memory Systems of Agentscited by 51
  7. 2025The FACTS Grounding Leaderboard: Benchmarking LLMs' Ability to Ground Responses to Long-Form Inputcited by 49
  8. 2025Large Language Models Hallucination: A Comprehensive Surveycited by 45
  9. 2026Skill Retrieval Augmentation for Agentic AIcited by 41
  10. 2024Detecting Hallucinations in Large Language Model Generation: A Token Probability Approachcited by 40
  11. 2025A Unified Understanding and Evaluation of Steering Methodscited by 40
  12. 2024A Probabilistic Framework for LLM Hallucination Detection via Belief Tree Propagationcited by 21
38 more
  1. 2025Efficient Hallucination Detection for LLMs Using Uncertainty-Aware Attention Headscited by 14
  2. 2025Depth Gives a False Sense of Privacy: LLM Internal States Inversioncited by 14
  3. 2025SurGE: A Benchmark and Evaluation Framework for Scientific Survey Generationcited by 14
  4. 2025Enhancing Hallucination Detection through Noise Injectioncited by 13
  5. 2024LeKUBE: A Legal Knowledge Update BEnchmarkcited by 12
  6. 2024FactCheckmate: Preemptively Detecting and Mitigating Hallucinations in LMscited by 11
  7. 2025EigenTrack: Spectral Activation Feature Tracking for Hallucination and Out-of-Distribution Detection in LLMs and VLMscited by 10
  8. 2025Attention Hijackers: Detect and Disentangle Attention Hijacking in LVLMs for Hallucination Mitigationcited by 9
  9. 2024The HalluRAG Dataset: Detecting Closed-Domain Hallucinations in RAG Applications Using an LLM's Internal Statescited by 8
  10. 2025Language Models Can Predict Their Own Behaviorcited by 8
  11. 2024A Survey of Automatic Hallucination Evaluation on Natural Language Generationcited by 7
  12. 2026TDGNet: Hallucination Detection in Diffusion Language Models via Temporal Dynamic Graphscited by 7
  13. 2025Loki's Dance of Illusions: A Comprehensive Survey of Hallucination in Large Language Modelscited by 6
  14. 2026FACTUM: Mechanistic Detection of Citation Hallucination in Long-Form RAGcited by 6
  15. 2026JailNewsBench: Multi-Lingual and Regional Benchmark for Fake News Generation under Jailbreak Attackscited by 6
  16. 2024AssistRAG: Boosting the Potential of Large Language Models with an Intelligent Information Assistantcited by 5
  17. 2024Beyond Introspection: Reinforcing Thinking via Externalist Behavioral Feedbackcited by 5
  18. 2025Uncertainty in Action: Confidence Elicitation in Embodied Agentscited by 5
  19. 2025Unsupervised Hallucination Detection by Inspecting Reasoning Processescited by 5
  20. 2025D2HScore: Reasoning-Aware Hallucination Detection via Semantic Breadth and Depth Analysis in LLMscited by 5
  21. 2024Layer Importance and Hallucination Analysis in Large Language Models via Enhanced Activation Variance-Sparsitycited by 4
  22. 2025Probing Neural Topology of Large Language Modelscited by 4
  23. 2025Bridging the Knowledge-Prediction Gap in LLMs on Multiple-Choice Questionscited by 4
  24. 2026Semantic Uncertainty Quantification of Hallucinations in LLMs: A Quantum Tensor Network Based Methodcited by 4
  25. 2026The Confidence Manifold: Geometric Structure of Correctness Representations in Language Modelscited by 4
  26. 2025Dynamic and Parametric Retrieval-Augmented Generationcited by 3
  27. 2026Architecting Trust in Artificial Epistemic Agentscited by 3
  28. 2026Atropos: Improving Cost-Benefit Trade-off of LLM-based Agents under Self-Consistency with Early Termination and Model Hotswapcited by 3
  29. 2026Low-Cost Black-Box Detection of LLM Hallucinations via Dynamical System Predictioncited by 3
  30. 2024Hallucination Detox: Sensitivity Dropout (SenD) for Large Language Model Trainingcited by 2
  31. 2025What are Models Thinking about? Understanding Large Language Model Hallucinations "Psychology" through Model Inner State Analysiscited by 2
  32. 2025From Hallucinations to Jailbreaks: Rethinking the Vulnerability of Large Foundation Modelscited by 2
  33. 2025High Accuracy, Less Talk (HALT): Reliable LLMs through Capability-Aligned Finetuningcited by 2
  34. 2025Decoding Memories: An Efficient Pipeline for Self-Consistency Hallucination Detectioncited by 2
  35. 2025Can Linear Probes Measure LLM Uncertainty?cited by 2
  36. 2026A Geometric Analysis of Small-sized Language Model Hallucinationscited by 2
  37. 2026Detection Without Correction: A Robust Asymmetry in Activation-Based Hallucination Probingcited by 2
  38. 2024LaMsS: When Large Language Models Meet Self-Skepticismcited by 1

50 most cited of 84 All 84 (JSON)