mlsanityData to 7 Oct 2026

Adversarial GLUE: A Multi-Task Benchmark for Robustness Evaluation of Language Models

arXiv 2111.02840 ↗First posted 4 Nov 2021cs.CL cs.CR cs.LGcited by 227

Not on the map. Why?

Cited by 31 in LLM safety and security, but placed by its own abstract, which didn’t fit.

Abstract (arXiv)

Large-scale pre-trained language models have achieved tremendous success across a wide range of natural language understanding (NLU) tasks, even surpassing human performance. However, recent studies reveal that the robustness of these models can be challenged by carefully crafted textual adversarial examples. While several individual datasets have been proposed to evaluate model robustness, a principled and comprehensive benchmark is still missing. In this paper, we present Adversarial GLUE (AdvGLUE), a new multi-task benchmark to quantitatively and thoroughly explore and evaluate the vulnerabilities of modern large-scale language models under various types of adversarial attacks. In particular, we systematically apply 14 textual adversarial attack methods to GLUE tasks to construct AdvGLUE, which is further validated by humans for reliable annotations. Our findings are summarized as follows. (i) Most existing adversarial attack algorithms are prone to generating invalid or ambiguous adversarial examples, with around 90% of them either changing the original semantic meanings or misleading human annotators as well. Therefore, we perform a careful filtering process to curate a high-quality benchmark. (ii) All the language models and robust training methods we tested perform poorly on AdvGLUE, with scores lagging far behind the benign accuracy. We hope our work will motivate the development of new adversarial attacks that are more stealthy and semantic-preserving, as well as new robust language models against sophisticated adversarial attacks. AdvGLUE is available at https://adversarialglue.github.io.

Key facts. Cited by 227 papers in the map (data through 7 Oct 2026). Builds on 52 of its 78 references that are in the map. First posted on arXiv 4 Nov 2021.

Builds on

52 of 78 references.

References · 52

  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. 2014Explaining and Harnessing Adversarial Examplescited by 7,287
  4. 2019BERTScore: Evaluating Text Generation with BERTcited by 5,462
  5. 2018GLUE: A Multi-Task Benchmark and Analysis Platform for Natural Language Understandingcited by 4,854
  6. 2016SQuAD: 100,000+ Questions for Machine Comprehension of Textcited by 4,115
  7. 2013Distributed Representations of Words and Phrases and their Compositionalitycited by 3,462
  8. 2019XLNet: Generalized Autoregressive Pretraining for Language Understandingcited by 2,801
  9. 2019ALBERT: A Lite BERT for Self-supervised Learning of Language Representationscited by 2,772
  10. 2017A Broad-Coverage Challenge Corpus for Sentence Understanding through Inferencecited by 2,665
  11. 2020DeBERTa: Decoding-enhanced BERT with Disentangled Attentioncited by 2,157
  12. 2015A large annotated corpus for learning natural language inferencecited by 1,837
38 more
  1. 2019SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systemscited by 1,718
  2. 2018Datasheets for Datasetscited by 1,371
  3. 2018FEVER: a large-scale dataset for Fact Extraction and VERificationcited by 1,362
  4. 2019Certified Adversarial Robustness via Randomized Smoothingcited by 1,324
  5. 2015DeepFool: a simple and accurate method to fool deep neural networkscited by 1,315
  6. 2020Beyond Accuracy: Behavioral Testing of NLP models with CheckListcited by 979
  7. 2019Adversarial NLI: A New Benchmark for Natural Language Understandingcited by 798
  8. 2015Distillation as a Defense to Adversarial Perturbations against Deep Neural Networkscited by 675
  9. 2019Is BERT Really Robust? A Strong Baseline for Natural Language Attack on Text Classification and Entailmentcited by 663
  10. 2017Adversarial Examples for Evaluating Reading Comprehension Systemscited by 568
  11. 2019ERNIE: Enhanced Language Representation with Informative Entitiescited by 525
  12. 2020BERT-ATTACK: Adversarial Attack Against BERT Using BERTcited by 488
  13. 2021Dynabench: Rethinking Benchmarking in NLPcited by 447
  14. 2018TextBugger: Generating Adversarial Text Against Real-world Applicationscited by 376
  15. 2020Large-Scale Adversarial Training for Vision-and-Language Representation Learningcited by 327
  16. 2020BAE: BERT-based Adversarial Examples for Text Classificationcited by 316
  17. 2018Adversarial Example Generation with Syntactically Controlled Paraphrase Networkscited by 282
  18. 2018Audio Adversarial Examples: Targeted Attacks on Speech-to-Textcited by 268
  19. 2019FreeLB: Enhanced Adversarial Training for Natural Language Understandingcited by 261
  20. 2019SMART: Robust and Efficient Fine-Tuning for Pre-trained Natural Language Models through Principled Regularized Optimizationcited by 253
  21. 2018Stress Test Evaluation for Natural Language Inferencecited by 211
  22. 2021What Will it Take to Fix Benchmarking in Natural Language Understanding?cited by 200
  23. 2019Certified Robustness to Adversarial Word Substitutionscited by 164
  24. 2020Beat the AI: Investigating Adversarial Human Annotation for Reading Comprehensioncited by 131
  25. 2020Adversarial Training for Large Neural Language Modelscited by 109
  26. 2021Robustness Gym: Unifying the NLP Evaluation Landscapecited by 107
  27. 2020InfoBERT: Improving Robustness of Language Models from An Information Theoretic Perspectivecited by 92
  28. 2019Achieving Verified Robustness to Symbol Substitutions via Interval Bound Propagationcited by 87
  29. 2020OpenAttack: An Open-source Textual Adversarial Attack Toolkitcited by 84
  30. 2020Reevaluating Adversarial Examples in Natural Languagecited by 83
  31. 2020SAFER: A Structure-free Approach for Certified Robustness to Adversarial Word Substitutionscited by 72
  32. 2017Robust Physical-World Attacks on Deep Learning Modelscited by 70
  33. 2018Training verified learners with learned verifierscited by 50
  34. 2019Word-level Textual Adversarial Attacking as Combinatorial Optimizationcited by 37
  35. 2018Characterizing Audio Adversarial Examples Using Temporal Dependencycited by 34
  36. 2019Human vs. Muppet: A Conservative Estimate of Human Performance on the GLUE Benchmarkcited by 32
  37. 2019OpenHowNet: An Open Sememe-based Lexical Knowledge Basecited by 21
  38. 2019Adversarial attacks against Fact Extraction and VERificationcited by 11

50 most cited of 52 All 52 (JSON)

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 1; 2022: 24; 2023: 53; 2024: 74; 2025: 48; 2026: 27.

227 papers.

Sort:

By area:

  1. 2023Universal and Transferable Adversarial Attacks on Aligned Language Modelscited by 2,914
  2. 2024HarmBench: A Standardized Evaluation Framework for Automated Red Teaming and Robust Refusalcited by 1,330
  3. 2023A Survey on Evaluation of Large Language Modelscited by 1,227
  4. 2023Large Language Models Can Be Easily Distracted by Irrelevant Contextcited by 820
  5. 2023XSTest: A Test Suite for Identifying Exaggerated Safety Behaviours in Large Language Modelscited by 477
  6. 2023State of What Art? A Call for Multi-Prompt LLM Evaluationcited by 299
  7. 2023On Evaluating Adversarial Robustness of Large Vision-Language Modelscited by 281
  8. 2022Prompting GPT-3 To Be Reliablecited by 258
  9. 2023From Pretraining Data to Language Models to Downstream Tasks: Tracking the Trails of Political Biases Leading to Unfair NLP Modelscited by 257
  10. 2024OR-Bench: An Over-Refusal Benchmark for Large Language Modelscited by 222
  11. 2023PromptRobust: Towards Evaluating the Robustness of Large Language Models on Adversarial Promptscited by 218
  12. 2023On the Robustness of ChatGPT: An Adversarial and Out-of-distribution Perspectivecited by 208
38 more
  1. 2023Better Diffusion Models Further Improve Adversarial Trainingcited by 207
  2. 2022Can large language models reason about medical questions?cited by 194
  3. 2024Political Compass or Spinning Arrow? Towards More Meaningful Evaluations for Values and Opinions in Large Language Modelscited by 162
  4. 2023Poisoning Retrieval Corpora by Injecting Adversarial Passagescited by 153
  5. 2024Benchmark Data Contamination of Large Language Models: A Surveycited by 138
  6. 2023Auditing large language models: a three-layered approachcited by 135
  7. 2024On Memorization of Large Language Models in Logical Reasoningcited by 114
  8. 2025A Comprehensive Survey in LLM(-Agent) Full Stack Safety: Data, Training and Deploymentcited by 110
  9. 2021Measure and Improve Robustness in NLP Models: A Surveycited by 107
  10. 2023Graph Foundation Models: Concepts, Opportunities and Challengescited by 101
  11. 2022ReCode: Robustness Evaluation of Code Generation Modelscited by 100
  12. 2023Revisiting Out-of-distribution Robustness in NLP: Benchmark, Analysis, and LLMs Evaluationscited by 96
  13. 2024ALERT: A Comprehensive Benchmark for Assessing Large Language Models' Safety through Red Teamingcited by 96
  14. 2023A Survey of Safety and Trustworthiness of Large Language Models through the Lens of Verification and Validationcited by 91
  15. 2022Shortcut Learning of Large Language Models in Natural Language Understandingcited by 83
  16. 2023A Pilot Study of Query-Free Adversarial Attack against Stable Diffusioncited by 75
  17. 2024Evaluating Quantized Large Language Modelscited by 75
  18. 2024BadChain: Backdoor Chain-of-Thought Prompting for Large Language Modelscited by 74
  19. 2022GLUE-X: Evaluating Natural Language Understanding Models from an Out-of-distribution Generalization Perspectivecited by 72
  20. 2025STAIR: Improving Safety Alignment with Introspective Reasoningcited by 72
  21. 2024When LLMs Meet Cybersecurity: A Systematic Literature Reviewcited by 70
  22. 2022Exploring the Limits of Domain-Adaptive Training for Detoxifying Large-Scale Language Modelscited by 67
  23. 2023VLATTACK: Multimodal Adversarial Attacks on Vision-Language Tasks via Pre-trained Modelscited by 66
  24. 2024Typos that Broke the RAG's Back: Genetic Attack on RAG Pipeline by Simulating Documents in the Wild via Low-level Perturbationscited by 64
  25. 2024Risk Taxonomy, Mitigation, and Assessment Benchmarks of Large Language Model Systemscited by 63
  26. 2022Adversarial Training for High-Stakes Reliabilitycited by 60
  27. 2023Adversarial Demonstration Attacks on Large Language Modelscited by 58
  28. 2024Safetywashing: Do AI Safety Benchmarks Actually Measure Safety Progress?cited by 56
  29. 2024Survey on Knowledge Distillation for Large Language Models: Methods, Evaluation, and Applicationcited by 55
  30. 2023Evaluating the Instruction-Following Robustness of Large Language Models to Prompt Injectioncited by 53
  31. 2023Dr.Spider: A Diagnostic Evaluation Benchmark towards Text-to-SQL Robustnesscited by 50
  32. 2024Operationalizing a Threat Model for Red-Teaming Large Language Models (LLMs)cited by 44
  33. 2024M4GT-Bench: Evaluation Benchmark for Black-Box Machine-Generated Text Detectioncited by 43
  34. 2022SemAttack: Natural Textual Attacks via Different Semantic Spacescited by 42
  35. 2023On Robustness of Prompt-based Semantic Parsing with Large Pre-trained Language Model: An Empirical Study on Codexcited by 41
  36. 2023RobuT: A Systematic Study of Table QA Robustness Against Human-Annotated Adversarial Perturbationscited by 41
  37. 2024Prompt Stealing Attacks Against Large Language Modelscited by 39
  38. 2023GPTAraEval: A Comprehensive Evaluation of ChatGPT on Arabic NLPcited by 37

50 most cited of 227 All 227 (JSON)