mlsanityData to 14 Oct 2026

Large Language Models Struggle to Learn Long-Tail Knowledge

arXiv 2211.08411 ↗First posted 15 Nov 2022cs.CL cs.LGcited by 458

Not on the map. Why?

Cited by 93 in Retrieval-augmented generation, but placed by its own abstract, which didn’t fit.

Abstract (arXiv)

The Internet contains a wealth of knowledge -- from the birthdays of historical figures to tutorials on how to code -- all of which may be learned by language models. However, while certain pieces of information are ubiquitous on the web, others appear extremely rarely. In this paper, we study the relationship between the knowledge memorized by large language models and the information in pre-training datasets scraped from the web. In particular, we show that a language model's ability to answer a fact-based question relates to how many documents associated with that question were seen during pre-training. We identify these relevant documents by entity linking pre-training datasets and counting documents that contain the same entities as a given question-answer pair. Our results demonstrate strong correlational and causal relationships between accuracy and relevant document count for numerous question answering datasets (e.g., TriviaQA), pre-training corpora (e.g., ROOTS), and model sizes (e.g., 176B parameters). Moreover, while larger models are better at learning long-tail knowledge, we estimate that today's models must be scaled by many orders of magnitude to reach competitive QA performance on questions with little support in the pre-training data. Finally, we show that retrieval-augmentation can reduce the dependence on relevant pre-training information, presenting a promising approach for capturing the long-tail.

Key facts. Cited by 458 papers in the map (data through 14 Oct 2026). Builds on 30 of its 42 references that are in the map. First posted on arXiv 15 Nov 2022.

Builds on

30 of 42 references.

References · 30

  1. 2017Attention Is All You Needcited by 60,361
  2. 2020Language Models are Few-Shot Learnerscited by 32,419
  3. 2019Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformercited by 14,599
  4. 2020Retrieval-Augmented Generation for Knowledge-Intensive NLP Taskscited by 9,297
  5. 2016SQuAD: 100,000+ Questions for Machine Comprehension of Textcited by 4,148
  6. 2020Dense Passage Retrieval for Open-Domain Question Answeringcited by 3,824
  7. 2017TriviaQA: A Large Scale Distantly Supervised Challenge Dataset for Reading Comprehensioncited by 2,955
  8. 2020The Pile: An 800GB Dataset of Diverse Text for Language Modelingcited by 2,411
  9. 2020Extracting Training Data from Large Language Modelscited by 2,189
  10. 2020REALM: Retrieval-Augmented Language Model Pre-Trainingcited by 2,141
  11. 2022BLOOM: A 176B-Parameter Open-Access Multilingual Language Modelcited by 1,984
  12. 2019Language Models as Knowledge Bases?cited by 1,936
18 more
  1. 2018The Secret Sharer: Evaluating and Testing Unintended Memorization in Neural Networkscited by 864
  2. 2020How Much Knowledge Can You Pack Into the Parameters of a Language Model?cited by 719
  3. 2021Deduplicating Training Data Makes Language Models Bettercited by 714
  4. 2022GPT-NeoX-20B: An Open-Source Autoregressive Language Modelcited by 713
  5. 2019Latent Retrieval for Weakly Supervised Open Domain Question Answeringcited by 695
  6. 2020Autoregressive Entity Retrievalcited by 334
  7. 2022Deduplicating Training Data Mitigates Privacy Risks in Language Modelscited by 314
  8. 2020Distilling Knowledge from Reader to Retriever for Question Answeringcited by 226
  9. 2020How Context Affects Language Models' Factual Predictionscited by 181
  10. 2022What Language Model Architecture and Pretraining Objective Work Best for Zero-Shot Generalization?cited by 150
  11. 2022Impact of Pretraining Term Frequencies on Few-Shot Reasoningcited by 142
  12. 2022Measuring Forgetting of Memorized Training Examplescited by 118
  13. 2020Entities as Experts: Sparse Memory Access with Entity Supervisioncited by 100
  14. 2022On the Effect of Pretraining Corpora on In-context Learning by a Large-scale Language Modelcited by 81
  15. 2021Frequency Effects on Syntactic Rule Learning in Transformerscited by 59
  16. 2022Measuring Causal Effects of Data Statistics on Language Model's `Factual' Predictionscited by 57
  17. 2022ORCA: Interpreting Prompted Language Models via Locating Supporting Data Evidence in the Ocean of Pretraining Datacited by 34
  18. 2021Challenges in Generalization in Open Domain Question Answeringcited by 30

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2022: 6; 2023: 71; 2024: 166; 2025: 112; 2026: 103.

458 papers.

Sort:

By area:

  1. 2023Lost in the Middle: How Language Models Use Long Contextscited by 3,455
  2. 2022Self-Instruct: Aligning Language Models with Self-Generated Instructionscited by 2,724
  3. 2023Retrieval-Augmented Generation for Large Language Models: A Surveycited by 2,212
  4. 2023Pythia: A Suite for Analyzing Large Language Models Across Training and Scalingcited by 1,786
  5. 2023A Survey on Hallucination in Large Language Models: Principles, Taxonomy, Challenges, and Open Questionscited by 1,601
  6. 2023The Rise and Potential of Large Language Model Based Agents: A Surveycited by 1,439
  7. 2023FActScore: Fine-grained Atomic Evaluation of Factual Precision in Long Form Text Generationcited by 1,282
  8. 2022When Not to Trust Language Models: Investigating Effectiveness of Parametric and Non-Parametric Memoriescited by 1,128
  9. 2023Extracting Training Data from Diffusion Modelscited by 732
  10. 2024RAPTOR: Recursive Abstractive Processing for Tree-Organized Retrievalcited by 532
  11. 2023Ragas: Automated Evaluation of Retrieval Augmented Generationcited by 418
  12. 2023The Reversal Curse: LLMs trained on "A is B" fail to learn "B is A"cited by 407
38 more
  1. 2024Retrieval-Augmented Generation for AI-Generated Content: A Surveycited by 354
  2. 2023Explainability for Large Language Models: A Surveycited by 333
  3. 2024Hallucination is Inevitable: An Innate Limitation of Large Language Modelscited by 332
  4. 2024Patchscopes: A Unifying Framework for Inspecting Hidden Representations of Language Modelscited by 235
  5. 2022Demystifying Prompts in Language Models via Perplexity Estimationcited by 216
  6. 2023Can Pre-trained Vision and Language Models Answer Visual Information-Seeking Questions?cited by 207
  7. 2023Survey on Factuality in Large Language Models: Knowledge, Retrieval and Domain-Specificitycited by 206
  8. 2023Fine-Tuning or Retrieval? Comparing Knowledge Injection in LLMscited by 178
  9. 2023Head-to-Tail: How Knowledgeable are Large Language Models (LLMs)? A.K.A. Will LLMs Replace Knowledge Graphs?cited by 168
  10. 2024AI Agents Under Threat: A Survey of Key Security Challenges and Future Pathwayscited by 127
  11. 2023Speak, Memory: An Archaeology of Books Known to ChatGPT/GPT-4cited by 125
  12. 2022A Survey of Deep Learning for Mathematical Reasoningcited by 124
  13. 2025MMSearch-R1: Incentivizing LMMs to Searchcited by 124
  14. 2023From Word Models to World Models: Translating from Natural Language to the Probabilistic Language of Thoughtcited by 119
  15. 2024Rethinking Interpretability in the Era of Large Language Modelscited by 111
  16. 2024Bias and Unfairness in Information Retrieval Systems: New Challenges in the LLM Eracited by 98
  17. 2024CRAG -- Comprehensive RAG Benchmarkcited by 97
  18. 2025A Sober Look at Progress in Language Model Reasoning: Pitfalls and Paths to Reproducibilitycited by 93
  19. 2023Language Model Behavior: A Comprehensive Surveycited by 92
  20. 2024RAG-Driver: Generalisable Driving Explanations with Retrieval-Augmented In-Context Learning in Multi-Modal Large Language Modelcited by 92
  21. 2025HalluLens: LLM Hallucination Benchmarkcited by 92
  22. 2023Diffusion Models for Reinforcement Learning: A Surveycited by 91
  23. 2024CodeRAG-Bench: Can Retrieval Augment Code Generation?cited by 91
  24. 2024How Do Large Language Models Acquire Factual Knowledge During Pretraining?cited by 90
  25. 2023Improving Language Models via Plug-and-Play Retrieval Feedbackcited by 87
  26. 2024Unfamiliar Finetuning Examples Control How Language Models Hallucinatecited by 87
  27. 2023Cognitive Mirage: A Review of Hallucinations in Large Language Modelscited by 84
  28. 2023The MiniPile Challenge for Data-Efficient Language Modelscited by 75
  29. 2024No "Zero-Shot" Without Exponential Data: Pretraining Concept Frequency Determines Multimodal Model Performancecited by 75
  30. 2023Search-in-the-Chain: Interactively Enhancing Large Language Models with Search for Knowledge-intensive Taskscited by 73
  31. 2024Reliable, Adaptable, and Attributable Language Models with Retrievalcited by 73
  32. 2023The Bias Amplification Paradox in Text-to-Image Generationcited by 70
  33. 2023Large Language Models can Learn Rulescited by 69
  34. 2023Can LMs Learn New Entities from Descriptions? Challenges in Propagating Injected Knowledgecited by 67
  35. 2023Attention Satisfies: A Constraint-Satisfaction Lens on Factual Errors of Language Modelscited by 67
  36. 2024Risk Taxonomy, Mitigation, and Assessment Benchmarks of Large Language Model Systemscited by 66
  37. 2025KVLink: Accelerating Large Language Models via Efficient KV Cache Reusecited by 66
  38. 2024Quantifying Contamination in Evaluating Code Generation Capabilities of Language Modelscited by 65

50 most cited of 458 All 458 (JSON)