mlsanityData to 7 Oct 2026

Memorization vs. Generalization: Quantifying Data Leakage in NLP Performance Evaluation

arXiv 2102.01818 ↗First posted 3 Feb 2021cs.CL cs.LGcited by 69

Not on the map. Why?

Abstract (arXiv)

Public datasets are often used to evaluate the efficacy and generalizability of state-of-the-art methods for many tasks in natural language processing (NLP). However, the presence of overlap between the train and test datasets can lead to inflated results, inadvertently evaluating the model's ability to memorize and interpreting it as the ability to generalize. In addition, such data sets may not provide an effective indicator of the performance of these methods in real world scenarios. We identify leakage of training data into test data on several publicly available datasets used to evaluate NLP tasks, including named entity recognition and relation extraction, and study them to assess the impact of that leakage on the model's ability to memorize versus generalize.

Key facts. Cited by 69 papers in the map (data through 7 Oct 2026). Builds on 9 of its 35 references that are in the map. First posted on arXiv 3 Feb 2021.

Builds on

9 of 35 references.

References · 9

  1. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,204
  2. 2018GLUE: A Multi-Task Benchmark and Analysis Platform for Natural Language Understandingcited by 4,854
  3. 2017TriviaQA: A Large Scale Distantly Supervised Challenge Dataset for Reading Comprehensioncited by 2,895
  4. 2019BioBERT: a pre-trained biomedical language representation model for biomedical text miningcited by 1,935
  5. 2017SemEval-2017 Task 1: Semantic Textual Similarity - Multilingual and Cross-lingual Focused Evaluationcited by 998
  6. 2020Question and Answer Test-Train Overlap in Open-Domain Question Answering Datasetscited by 137
  7. 2018emrQA: A Large Corpus for Question Answering on Electronic Medical Recordscited by 134
  8. 2018MedSTS: A Resource for Clinical Semantic Textual Similaritycited by 39
  9. 2020Entity-Enriched Neural Models for Clinical Question Answeringcited by 8

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 4; 2022: 10; 2023: 15; 2024: 22; 2025: 11; 2026: 7.

69 papers.

Sort:

By area:

  1. 2022DS-1000: A Natural and Reliable Benchmark for Data Science Code Generationcited by 442
  2. 2023NLP Evaluation in trouble: On the Need to Measure LLM Data Contamination for each Benchmarkcited by 339
  3. 2023Multilingual Machine Translation with Large Language Models: Empirical Results and Analysiscited by 156
  4. 2022Impact of Pretraining Term Frequencies on Few-Shot Reasoningcited by 142
  5. 2024Benchmark Data Contamination of Large Language Models: A Surveycited by 138
  6. 2023Speak, Memory: An Archaeology of Books Known to ChatGPT/GPT-4cited by 123
  7. 2025Preference Leakage: A Contamination Problem in LLM-as-a-judgecited by 119
  8. 2023Copyright Violations and Large Language Modelscited by 108
  9. 2024SpreadsheetBench: Towards Challenging Real World Spreadsheet Manipulationcited by 100
  10. 2022State-of-the-art generalisation research in NLP: A taxonomy and reviewcited by 97
  11. 2023CoMPosT: Characterizing and Evaluating Caricature in LLM Simulationscited by 94
  12. 2023The Data Provenance Initiative: A Large Scale Audit of Dataset Licensing & Attribution in AIcited by 73
38 more
  1. 2024Reasoning Paths with Reference Objects Elicit Quantitative Spatial Reasoning in Large Vision-Language Modelscited by 54
  2. 2024Consent in Crisis: The Rapid Decline of the AI Data Commonscited by 45
  3. 2022Decoupling Knowledge from Memorization: Retrieval-augmented Prompt Learningcited by 36
  4. 2022BigBIO: A Framework for Data-Centric Biomedical Natural Language Processingcited by 36
  5. 2023ALCUNA: Large Language Models Meet New Knowledgecited by 35
  6. 2025A Survey on Data Contamination for Large Language Modelscited by 33
  7. 2024Graph Transformers: A Surveycited by 32
  8. 2024Undesirable Memorization in Large Language Models: A Surveycited by 29
  9. 2024Does Data Contamination Detection Work (Well) for LLMs? A Survey and Evaluation on Detection Assumptionscited by 26
  10. 2024MERA: A Comprehensive LLM Evaluation in Russiancited by 22
  11. 2022An Empirical Study of Memorization in NLPcited by 21
  12. 2024Improving Model Evaluation using SMART Filtering of Benchmark Datasetscited by 20
  13. 2024When does compositional structure yield compositional generalization? A kernel theorycited by 19
  14. 2021Memorization and Generalization in Neural Code Intelligence Modelscited by 17
  15. 2023Exploring the Robustness of Large Language Models for Solving Programming Problemscited by 16
  16. 2024Language model developers should report train-test overlapcited by 16
  17. 2022Vote'n'Rank: Revision of Benchmarking with Social Choice Theorycited by 15
  18. 2021Cut the CARP: Fishing for zero-shot story evaluationcited by 14
  19. 2024Automated data processing and feature engineering for deep learning and big data applications: a surveycited by 14
  20. 2024ONEBench to Test Them All: Sample-Level Benchmarking Over Open-Ended Capabilitiescited by 13
  21. 2024A Comprehensive Survey of Contamination Detection Methods in Large Language Modelscited by 12
  22. 2022TAPE: Assessing Few-shot Russian Language Understandingcited by 11
  23. 2024DICE: Detecting In-distribution Contamination in LLM's Fine-tuning Phase for Math Reasoningcited by 11
  24. 2021How not to Lie with a Benchmark: Rearranging NLP Leaderboardscited by 9
  25. 2024BLAZE: Cross-Language and Cross-Project Bug Localization via Dynamic Chunking and Hard Example Learningcited by 7
  26. 2023Battle of the Large Language Models: Dolly vs LLaMA vs Vicuna vs Guanaco vs Bard vs ChatGPT -- A Text-to-SQL Parsing Comparisoncited by 6
  27. 2024On the Workflows and Smells of Leaderboard Operations (LBOps): An Exploratory Study of Foundation Model Leaderboardscited by 6
  28. 2021How Emotionally Stable is ALBERT? Testing Robustness with Stochastic Weight Averaging on a Sentiment Analysis Taskcited by 5
  29. 2022Hollywood Identity Bias Dataset: A Context Oriented Bias Analysis of Movie Dialoguescited by 5
  30. 2023On Evaluation of Document Classification using RVL-CDIPcited by 5
  31. 2024SpanSeq: Similarity-based sequence data splitting method for improved development and assessment of deep learning projectscited by 5
  32. 2025Memorization in Large Language Models in Medicine: Prevalence, Characteristics, and Implicationscited by 5
  33. 2023Principles from Clinical Research for NLP Model Generalizationcited by 4
  34. 2025Deep Contrastive Unlearning for Language Modelscited by 4
  35. 2025SUV: Scalable Large Language Model Copyright Compliance with Regularized Selective Unlearningcited by 4
  36. 2023Effects of Human Adversarial and Affable Samples on BERT Generalizationcited by 3
  37. 2024Scalability of Bayesian Network Structure Elicitation with Large Language Models: a Novel Methodology and Comparative Analysiscited by 3
  38. 2025OWL: Probing Cross-Lingual Recall of Memorized Texts via World Literaturecited by 3

50 most cited of 69 All 69 (JSON)