mlsanityData to 7 Oct 2026

Do Concept Bottleneck Models Learn as Intended?

arXiv 2105.04289 ↗First posted 10 May 2021cs.LG cs.AIcited by 114

Not on the map. Why?

Abstract (arXiv)

Concept bottleneck models map from raw inputs to concepts, and then from concepts to targets. Such models aim to incorporate pre-specified, high-level concepts into the learning procedure, and have been motivated to meet three desiderata: interpretability, predictability, and intervenability. However, we find that concept bottleneck models struggle to meet these goals. Using post hoc interpretability methods, we demonstrate that concepts do not correspond to anything semantically meaningful in input space, thus calling into question the usefulness of concept bottleneck models in their current form.

Key facts. Cited by 114 papers in the map (data through 7 Oct 2026). Builds on 6 of its 11 references that are in the map. First posted on arXiv 10 May 2021.

Builds on

6 of 11 references.

References · 6

  1. 2016"Why Should I Trust You?": Explaining the Predictions of Any Classifiercited by 5,040
  2. 2017Axiomatic Attribution for Deep Networkscited by 3,350
  3. 2017SmoothGrad: removing noise by adding noisecited by 1,097
  4. 2020Concept Bottleneck Modelscited by 1,022
  5. 2020Captum: A unified and generic model interpretability library for PyTorchcited by 542
  6. 2019Explainable Machine Learning in Deploymentcited by 210

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 3; 2022: 13; 2023: 16; 2024: 34; 2025: 25; 2026: 23.

114 papers.

Sort:

By area:

  1. 2022Language in a Bottle: Language Model Guided Concept Bottlenecks for Interpretable Image Classificationcited by 310
  2. 2022Post-hoc Concept Bottleneck Modelscited by 289
  3. 2021Promises and Pitfalls of Black-Box Concept Learning Modelscited by 128
  4. 2024Interpreting CLIP with Sparse Linear Concept Embeddings (SpLiCE)cited by 116
  5. 2021HIVE: Evaluating the Human Interpretability of Visual Explanationscited by 105
  6. 2023Concept-based Explainable Artificial Intelligence: A Surveycited by 97
  7. 2024Discover-then-Name: Task-Agnostic Concept Bottlenecks via Automated Concept Discoverycited by 91
  8. 2021Transparency of Deep Neural Networks for Medical Image Analysis: A Review of Interpretability Methodscited by 83
  9. 2022Concept Bottleneck Model with Additional Unsupervised Conceptscited by 79
  10. 2022GlanceNets: Interpretabile, Leak-proof Concept-based Modelscited by 73
  11. 2024VLG-CBM: Training Concept Bottleneck Models with Vision-Language Guidancecited by 65
  12. 2022"Help Me Help the AI": Understanding How Explainability Can Support Human-AI Interactioncited by 63
38 more
  1. 2023A Closer Look at the Intervention Procedure of Concept Bottleneck Modelscited by 63
  2. 2022Concept Activation Regions: A Generalized Framework For Concept-Based Explanationscited by 57
  3. 2022Leveraging Explanations in Interactive Machine Learning: An Overviewcited by 52
  4. 2024Stochastic Concept Bottleneck Modelscited by 49
  5. 2022Overlooked factors in concept-based explanations: Dataset choice, concept learnability, and human capabilitycited by 46
  6. 2024Incremental Residual Concept Bottleneck Modelscited by 43
  7. 2023Human Uncertainty in Concept-Based AI Systemscited by 39
  8. 2023Auxiliary Losses for Learning Generalizable Concept-based Modelscited by 35
  9. 2023Towards Robust Metrics for Concept Representation Evaluationcited by 29
  10. 2023Learning to Intervene on Concept Bottleneckscited by 29
  11. 2024Beyond Concept Bottleneck Models: How to Make Black Boxes Intervenable?cited by 28
  12. 2023SI-MIL: Taming Deep MIL for Self-Interpretability in Gigapixel Histopathologycited by 25
  13. 2022Human-Centered Concept Explanations for Neural Networkscited by 23
  14. 2022Understanding and Enhancing Robustness of Concept-based Modelscited by 22
  15. 2023Interpretability is in the Mind of the Beholder: A Causal Framework for Human-interpretable Representation Learningcited by 22
  16. 2024On the Concept Trustworthiness in Concept Bottleneck Modelscited by 22
  17. 2024Concept Bottleneck Models Without Predefined Conceptscited by 17
  18. 2024Concept-Based Explanations in Computer Vision: Where Are We and Where Could We Go?cited by 17
  19. 2024Pre-trained Vision-Language Models Learn Discoverable Visual Conceptscited by 16
  20. 2024Bayesian Concept Bottleneck Models with LLM Priorscited by 16
  21. 2024Concept Bottleneck Language Models For protein designcited by 15
  22. 2025If Concept Bottlenecks are the Question, are Foundation Models the Answer?cited by 15
  23. 2025Shortcuts and Identifiability in Concept-based Models from a Neuro-Symbolic Lenscited by 13
  24. 2025Leakage and Interpretability in Concept-Based Modelscited by 13
  25. 2023Cross-Modal Conceptualization in Bottleneck Modelscited by 12
  26. 2025Discovering Fine-Grained Visual-Concept Relations by Disentangled Optimal Transport Concept Bottleneck Modelscited by 12
  27. 2023Interpretable and intervenable ultrasonography-based machine learning models for pediatric appendicitiscited by 11
  28. 2023Evaluating the Stability of Semantic Concept Representations in CNNs for Robust Explainabilitycited by 11
  29. 2024Do Concept Bottleneck Models Respect Localities?cited by 10
  30. 2025Beyond Topological Self-Explainable GNNs: A Formal Explainability Perspectivecited by 10
  31. 2023Q-SENN: Quantized Self-Explaining Neural Networkscited by 9
  32. 2025Measuring Leakage in Concept-Based Methods: An Information Theoretic Approachcited by 9
  33. 2024A survey on Concept-based Approaches For Model Improvementcited by 8
  34. 2024Tree-Based Leakage Inspection and Control in Concept Bottleneck Modelscited by 8
  35. 2023Towards a Deeper Understanding of Concept Bottleneck Models Through End-to-End Explanationcited by 7
  36. 2023Take 5: Interpretable Image Classification with a Handful of Featurescited by 7
  37. 2025There Was Never a Bottleneck in Concept Bottleneck Modelscited by 7
  38. 2022Towards learning to explain with concept bottleneck models: mitigating information leakagecited by 6

50 most cited of 114 All 114 (JSON)