mlsanityData to 7 Oct 2026

Intrinsic Dimension Estimation for Robust Detection of AI-Generated Texts

arXiv 2306.04723 ↗First posted 7 Jun 2023cs.CL cs.AI cs.ITcited by 111

Not on the map. Why?

Abstract (arXiv)

Rapidly increasing quality of AI-generated content makes it difficult to distinguish between human and AI-generated texts, which may lead to undesirable consequences for society. Therefore, it becomes increasingly important to study the properties of human texts that are invariant over different text domains and varying proficiency of human writers, can be easily calculated for any language, and can robustly separate natural and AI-generated texts regardless of the generation model and sampling method. In this work, we propose such an invariant for human-written texts, namely the intrinsic dimensionality of the manifold underlying the set of embeddings for a given text sample. We show that the average intrinsic dimensionality of fluent texts in a natural language is hovering around the value 9 for several alphabet-based languages and around 7 for Chinese, while the average intrinsic dimensionality of AI-generated texts for each language is ≈ 1.5 lower, with a clear statistical separation between human-generated and AI-generated distributions. This property allows us to build a score-based artificial text detector. The proposed detector's accuracy is stable over text domains, generator models, and human writer proficiency levels, outperforming SOTA detectors in model-agnostic and cross-domain scenarios by a significant margin.

Key facts. Cited by 111 papers in the map (data through 7 Oct 2026). Builds on 41 of its 61 references that are in the map. First posted on arXiv 7 Jun 2023.

Builds on

41 of 61 references.

References · 41

  1. 2017Attention Is All You Needcited by 59,932
  2. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,204
  3. 2019RoBERTa: A Robustly Optimized BERT Pretraining Approachcited by 13,266
  4. 2021HuBERT: Self-Supervised Speech Representation Learning by Masked Prediction of Hidden Unitscited by 2,301
  5. 2018Hierarchical Neural Story Generationcited by 1,182
  6. 2023DetectGPT: Zero-Shot Machine-Generated Text Detection using Probability Curvaturecited by 705
  7. 2023A Watermark for Large Language Modelscited by 667
  8. 2020Intrinsic Dimensionality Explains the Effectiveness of Language Model Fine-Tuningcited by 658
  9. 2019Defending Against Neural Fake Newscited by 553
  10. 2019Release Strategies and the Social Impacts of Language Modelscited by 505
  11. 2019GLTR: Statistical Detection and Visualization of Generated Textcited by 490
  12. 2023How Close is ChatGPT to Human Experts? Comparison Corpus, Evaluation, and Detectioncited by 472
29 more
  1. 2023Paraphrasing evades detectors of AI-generated text, but retrieval is an effective defensecited by 401
  2. 2023Can AI-Generated Text be Reliably Detected?cited by 372
  3. 2018Measuring the Intrinsic Dimension of Objective Landscapescited by 364
  4. 2021The Intrinsic Dimension of Images and Its Impact on Learningcited by 307
  5. 2019Intrinsic dimension of data representations in deep neural networkscited by 276
  6. 2018Estimating the intrinsic dimension of datasets by a minimal neighborhood informationcited by 246
  7. 2023GPT detectors are biased against non-native English writerscited by 175
  8. 2017An introduction to Topological Data Analysis: fundamental and practical aspects for data scientistscited by 171
  9. 2020TweepFake: about Detecting Deepfake Tweetscited by 117
  10. 2019Ripser: efficient computation of Vietoris-Rips persistence barcodescited by 115
  11. 2023The geometry of hidden representations of large transformer modelscited by 108
  12. 2019Real or Fake? Learning to Discriminate Machine from Human Generated Textcited by 98
  13. 2023MGTBench: Benchmarking Machine-Generated Text Detectioncited by 90
  14. 2021Larger-Scale Transformers for Multilingual Masked Language Modelingcited by 86
  15. 2022RankGen: Improving Text Generation with Large Ranking Modelscited by 72
  16. 2021Intrinsic Dimension, Persistent Homology and Generalization in Neural Networkscited by 70
  17. 2021Scikit-dimension: a Python package for intrinsic dimension estimationcited by 68
  18. 2019Generating Sentiment-Preserving Fake Online Reviews Using Neural Language Models and Their Human- and Machine-based Detectioncited by 50
  19. 2021Artificial Text Detection via Examining the Topology of Attention Mapscited by 49
  20. 2021Manifold Topology Divergence: a Framework for Comparing Data Manifoldscited by 31
  21. 2022Intrinsic Dimensionality Estimation within Tight Localities: A Theoretical and Experimental Analysiscited by 22
  22. 2017LDMNet: Low Dimensional Manifold Regularized Neural Networkscited by 21
  23. 2019Estimating the effective dimension of large biological datasets using Fisher separability analysiscited by 19
  24. 2022Acceptability Judgements via Examining the Topology of Attention Mapscited by 19
  25. 2018Persistent Homology and the Upper Box Dimensioncited by 15
  26. 2022Topology and geometry of data manifold in deep learningcited by 12
  27. 2022Topological Data Analysis for Speech Processingcited by 9
  28. 2022Searching Intrinsic Dimensions of Vision Transformerscited by 2
  29. 2017Machine Learning in Appearance-based Robot Self-localizationcited by 1

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2023: 10; 2024: 40; 2025: 35; 2026: 26.

111 papers.

Sort:

By area:

  1. 2024Spotting LLMs With Binoculars: Zero-Shot Detection of Machine-Generated Textcited by 255
  2. 2024Monitoring AI-Modified Content at Scale: A Case Study on the Impact of ChatGPT on AI Conference Peer Reviewscited by 160
  3. 2024Mapping the Increasing Use of LLMs in Scientific Paperscited by 112
  4. 2023A Survey on LLM-Generated Text Detection: Necessity, Methods, and Future Directionscited by 101
  5. 2024Detecting Multimedia Generated by Large AI Models: A Surveycited by 73
  6. 2023A Survey on Detection of LLMs-Generated Contentcited by 70
  7. 2024Position: Topological Deep Learning is the New Frontier for Relational Learningcited by 65
  8. 2024Emergence of a High-Dimensional Abstraction Phase in Language Transformerscited by 56
  9. 2024Characterizing Truthfulness in Large Language Model Generations with Local Intrinsic Dimensioncited by 45
  10. 2024A Geometric View of Data Complexity: Efficient Local Intrinsic Dimension Estimation with Diffusion Modelscited by 41
  11. 2023The Shape of Learning: Anisotropy and Intrinsic Dimensions in Transformer-Based Modelscited by 37
  12. 2024GumbelSoft: Diversified Language Model Watermarking via the GumbelMax-trickcited by 31
38 more
  1. 2024Detecting AI-Generated Text: Factors Influencing Detectability with Current Methodscited by 23
  2. 2023AI-generated text boundary detection with RoFTcited by 20
  3. 2025AdaDetectGPT: Adaptive Detection of LLM-Generated Text with Statistical Guaranteescited by 20
  4. 2023Evade ChatGPT Detectors via A Single Spacecited by 18
  5. 2024Robust AI-Generated Text Detection by Restricted Embeddingscited by 18
  6. 2025Hallucination Detection in LLMs with Topological Divergence on Attention Graphscited by 18
  7. 2024Permute-and-Flip: An optimally stable and watermarkable decoder for LLMscited by 17
  8. 2025The Intrinsic Dimension of Prompts in Internal Representations of Large Language Modelscited by 17
  9. 2024Are AI-Generated Text Detectors Robust to Adversarial Perturbations?cited by 16
  10. 2023On the Trustworthiness Landscape of State-of-the-art Generative Models: A Survey and Outlookcited by 15
  11. 2023Counter Turing Test CT^2: AI-Generated Text Detection is Not as Easy as You May Think -- Introducing AI Detectability Indexcited by 14
  12. 2024Geometric Signatures of Compositionality Across a Language Model's Lifetimecited by 14
  13. 2024Glimpse: Enabling White-Box Methods to Use Proprietary Models for Zero-Shot LLM-Generated Text Detectioncited by 14
  14. 2025Feature-Level Insights into Artificial Text Detection with Sparse Autoencoderscited by 14
  15. 2025Token embeddings violate the manifold hypothesiscited by 14
  16. 2024DALD: Improving Logits-based Detector without Logits from Black-box LLMscited by 12
  17. 2025DETree: DEtecting Human-AI Collaborative Texts via Tree-Structured Hierarchical Representation Learningcited by 12
  18. 2024Is Contrasting All You Need? Contrastive Learning for the Detection and Attribution of AI-generated Textcited by 11
  19. 2024Persistent Topological Features in Large Language Modelscited by 11
  20. 2024A Practical Examination of AI-Generated Text Detectors for Large Language Modelscited by 11
  21. 2025GenAI Content Detection Task 1: English and Multilingual Machine-Generated Text Detection: AI vs. Humancited by 11
  22. 2024The structure of the token space for large language modelscited by 10
  23. 2025Redundancy, Isotropy, and Intrinsic Dimensionality of Prompt-based Text Embeddingscited by 10
  24. 2025Measuring the Intrinsic Dimension of Earth Representationscited by 10
  25. 2026GPTZero: Robust Detection of LLM-Generated Textscited by 10
  26. 2023Who Said That? Benchmarking Social Media AI Detectioncited by 9
  27. 2025textttDroid: A Resource Suite for AI-Generated Code Detectioncited by 9
  28. 2024'Quis custodiet ipsos custodes?' Who will watch the watchmen? On Detecting AI-generated peer-reviewscited by 8
  29. 2024Are AI Detectors Good Enough? A Survey on Quality of Datasets With Machine-Generated Textscited by 8
  30. 2025OpenTuringBench: An Open-Model-based Benchmark and Framework for Machine-Generated Text Detection and Attributioncited by 8
  31. 2025Human Texts Are Outliers: Detecting LLM-generated Texts via Out-of-distribution Detectioncited by 8
  32. 2025Geometry of Decision Making in Language Modelscited by 8
  33. 2026Detecting LLM-Generated Text with Performance Guaranteescited by 8
  34. 2026Learn-to-Distance: Distance Learning for Detecting LLM-Generated Textcited by 8
  35. 2025Authorship Attribution in Multilingual Machine-Generated Textscited by 7
  36. 2024The Impact of Prompts on Zero-Shot Detection of AI-Generated Textcited by 6
  37. 2024Humanizing the Machine: Proxy Attacks to Mislead LLM Detectorscited by 6
  38. 2025On the Predictive Power of Representation Dispersion in Language Modelscited by 6

50 most cited of 111 All 111 (JSON)