mlsanityData to 7 Oct 2026

How Pre-trained Language Models Capture Factual Knowledge? A Causal-Inspired Analysis

arXiv 2203.16747 ↗First posted 31 Mar 2022cs.CLcited by 42

Not on the map. Why?

Abstract (arXiv)

Recently, there has been a trend to investigate the factual knowledge captured by Pre-trained Language Models (PLMs). Many works show the PLMs' ability to fill in the missing factual words in cloze-style prompts such as "Dante was born in [MASK]." However, it is still a mystery how PLMs generate the results correctly: relying on effective clues or shortcut patterns? We try to answer this question by a causal-inspired analysis that quantitatively measures and evaluates the word-level patterns that PLMs depend on to generate the missing words. We check the words that have three typical associations with the missing words: knowledge-dependent, positionally close, and highly co-occurred. Our analysis shows: (1) PLMs generate the missing factual words more by the positionally close and highly co-occurred words than the knowledge-dependent words; (2) the dependence on the knowledge-dependent words is more effective than the positionally close and highly co-occurred words. Accordingly, we conclude that the PLMs capture the factual knowledge ineffectively because of depending on the inadequate associations.

Key facts. Cited by 42 papers in the map (data through 7 Oct 2026). Builds on 16 of its 34 references that are in the map. First posted on arXiv 31 Mar 2022.

Builds on

16 of 34 references.

References · 16

  1. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,204
  2. 2019RoBERTa: A Robustly Optimized BERT Pretraining Approachcited by 13,266
  3. 2019ALBERT: A Lite BERT for Self-supervised Learning of Language Representationscited by 2,772
  4. 2019Language Models as Knowledge Bases?cited by 1,922
  5. 2016ConceptNet 5.5: An Open Multilingual Graph of General Knowledgecited by 1,065
  6. 2019How Can We Know What Language Models Know?cited by 953
  7. 2019SpanBERT: Improving Pre-training by Representing and Predicting Spanscited by 832
  8. 2021Measuring and Improving Consistency in Pretrained Language Modelscited by 361
  9. 2021Factual Probing Is [MASK]: Learning vs. Learning to Recallcited by 291
  10. 2016Understanding Neural Networks through Representation Erasurecited by 223
  11. 2021Causal Inference in Natural Language Processing: Estimation, Prediction, Interpretation and Beyondcited by 177
  12. 2021Multilingual LAMA: Investigating Knowledge in Multilingual Pretrained Language Modelscited by 132
4 more
  1. 2020CausaLM: Causal Model Explanation Through Counterfactual Language Modelscited by 117
  2. 2021Knowledgeable or Educated Guess? Revisiting Language Models as Knowledge Basescited by 107
  3. 2019E-BERT: Efficient-Yet-Effective Entity Embeddings for BERTcited by 96
  4. 2021Interpreting Deep Learning Models in Natural Language Processing: A Reviewcited by 19

Built on by

Cited by year

Citing papers per year: 2022: 2; 2023: 17; 2024: 15; 2025: 8.

42 papers.

Sort:

By area:

  1. 2023A Survey on Hallucination in Large Language Models: Principles, Taxonomy, Challenges, and Open Questionscited by 1,582
  2. 2023Siren's Song in the AI Ocean: A Survey on Hallucination in Large Language Modelscited by 695
  3. 2023Unifying Large Language Models and Knowledge Graphs: A Roadmapcited by 663
  4. 2024Knowledge Conflicts for LLMs: A Surveycited by 274
  5. 2025Why Language Models Hallucinatecited by 205
  6. 2022Towards Faithful Model Explanation in NLP: A Surveycited by 129
  7. 2023Generating Benchmarks for Factuality Evaluation of Language Modelscited by 103
  8. 2023Graph Foundation Models: Concepts, Opportunities and Challengescited by 101
  9. 2024Bias and Unfairness in Information Retrieval Systems: New Challenges in the LLM Eracited by 97
  10. 2024The Dawn After the Dark: An Empirical Study on Factuality Hallucination in Large Language Modelscited by 87
  11. 2024How Do Large Language Models Acquire Factual Knowledge During Pretraining?cited by 86
  12. 2022A Survey of Knowledge Enhanced Pre-trained Language Modelscited by 76
30 more
  1. 2024Risk Taxonomy, Mitigation, and Assessment Benchmarks of Large Language Model Systemscited by 63
  2. 2023Journey to the Center of the Knowledge Neurons: Discoveries of Language-Independent Knowledge Neurons and Degenerate Knowledge Neuronscited by 56
  3. 2024Knowledge Mechanisms in Large Language Models: A Survey and Perspectivecited by 50
  4. 2023"According to ...": Prompting Language Models Improves Quoting from Pre-Training Datacited by 42
  5. 2023Give Me the Facts! A Survey on Factual Knowledge Probing in Pre-trained Language Modelscited by 32
  6. 2024On Behalf of the Stakeholders: Trends in NLP Model Interpretability in the Era of LLMscited by 24
  7. 2023The Life Cycle of Knowledge in Big Language Models: A Surveycited by 23
  8. 2023Impact of Co-occurrence on Factual Knowledge of Large Language Modelscited by 19
  9. 2024InfuserKI: Enhancing Large Language Models with Knowledge Graphs via Infuser-Guided Knowledge Integrationcited by 19
  10. 2025BiasEdit: Debiasing Stereotyped Language Models via Model Editingcited by 18
  11. 2023Data Management For Training Large Language Models: A Surveycited by 17
  12. 2023On the Trustworthiness Landscape of State-of-the-art Generative Models: A Survey and Outlookcited by 15
  13. 2023A Review of the Role of Causality in Developing Trustworthy AI Systemscited by 14
  14. 2025Why and How LLMs Hallucinate: Connecting the Dots with Subsequence Associationscited by 14
  15. 2024Look Within, Why LLMs Hallucinate: A Causal Perspectivecited by 11
  16. 2023Causality-aware Concept Extraction based on Knowledge-guided Promptingcited by 8
  17. 2025Robust Data Watermarking in Language Models by Injecting Fictitious Knowledgecited by 8
  18. 2024The Queen of England is not England's Queen: On the Lack of Factual Coherency in PLMscited by 7
  19. 2024ProbGate at EHRSQL 2024: Enhancing SQL Query Generation Accuracy through Probabilistic Threshold Filtering and Error Handlingcited by 6
  20. 2023Images in Language Space: Exploring the Suitability of Large Language Models for Vision & Language Taskscited by 5
  21. 2024IPL: Leveraging Multimodal Large Language Models for Intelligent Product Listingcited by 5
  22. 2024Entropy Guided Extrapolative Decoding to Improve Factuality in Large Language Modelscited by 4
  23. 2023Towards Faithful Knowledge Graph Explanation Through Deep Alignment in Commonsense Question Answeringcited by 3
  24. 2025Facts in Stats: Impacts of Pretraining Diversity on Language Model Generalizationcited by 3
  25. 2025When Bias Pretends to Be Truth: How Spurious Correlations Undermine Hallucination Detection in LLMscited by 3
  26. 2025Evaluating the Unseen Capabilities: How Many Theorems Do LLMs Know?cited by 2
  27. 2024A Comprehensive Evaluation of Semantic Relation Knowledge of Pretrained Language Models and Humanscited by 1
  28. 2025KSHSeek: Data-Driven Approaches to Mitigating and Detecting Knowledge-Shortcut Hallucinations in Generative Modelscited by 1
  29. 2023Examining the Emergence of Deductive Reasoning in Generative Language Modelscited by 0
  30. 2024BayesPrompt: Prompting Large-Scale Pre-Trained Language Models on Few-shot Inference via Debiased Domain Abstractioncited by 0