Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data | Paper Citation Record

Observation 7dd8691d-a4c7-4c72-8b0c-96dfb5ff8454 · outbound

This paper cites From attribution maps to human- understandable explanations through concept relevance propagation.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data From attribution maps to human- understandable explanations through concept relevance propagation

Reference 1

Resolution: unresolved; no resolver link, observed 2026-08-10T15:39:37.371501Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 3c3e6011-9a44-40e3-9f5f-1a527c653315 · outbound

This paper cites Under- standing intermediate layers using linear classi- fier probes.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Under- standing intermediate layers using linear classi- fier probes

Reference 2

Resolution: unresolved; no resolver link, observed 2026-08-10T15:39:37.375769Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 94b16547-f112-4759-9917-1cb60e3ac57e · outbound

This paper cites Finding and removing clever hans: Using explanation meth- ods to debug and improve deep models.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Finding and removing clever hans: Using explanation meth- ods to debug and improve deep models

Reference 3

Resolution: unresolved; no resolver link, observed 2026-08-10T15:39:37.379510Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation f7b50734-df47-4c52-9857-48244f487a4c · outbound

This paper cites On pixel-wise ex- planations for non-linear classifier decisions by layer-wise relevance propagation.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data On pixel-wise ex- planations for non-linear classifier decisions by layer-wise relevance propagation

Reference 4

Resolution: unresolved; no resolver link, observed 2026-08-10T15:39:37.383535Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation f244cbd4-e706-4610-bd75-0af60759049c · outbound

This paper cites Reactive model correction: Mitigating harm to task-relevant features via conditional bias suppression.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Reactive model correction: Mitigating harm to task-relevant features via conditional bias suppression

Reference 5

Resolution: unresolved; no resolver link, observed 2026-08-10T15:39:37.387339Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation ae32d32f-7dab-4f9d-ac9c-3d2f96bb25a7 · outbound

This paper cites Understanding the role of individual units in a deep neural network.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Understanding the role of individual units in a deep neural network

Reference 6

Resolution: unresolved; no resolver link, observed 2026-08-10T15:39:37.390953Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 1486d5c1-3060-4dbe-a4e0-24e4b76985f1 · outbound

This paper cites Explainability for fair machine learning.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Explainability for fair machine learning

Reference 7

Resolution: unresolved; no resolver link, observed 2026-08-10T15:39:37.394940Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 33fb3f44-42bc-4343-a222-e69b68b0dc54 · outbound

This paper cites Probing classifiers: Promises, shortcomings, and advances.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Probing classifiers: Promises, shortcomings, and advances

Reference 8

Resolution: unresolved; no resolver link, observed 2026-08-10T15:39:37.398992Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 644fe819-8db5-402e-9f40-555dc19b4bb7 · outbound

This paper cites Leace: Perfect linear concept erasure in closed form.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Leace: Perfect linear concept erasure in closed form

Reference 9

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:38.650703Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation 56cdf558-b861-4989-8c6d-95ea4f5efcaa · outbound

This paper cites Debiasing skin lesion datasets and models? not so fast.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Debiasing skin lesion datasets and models? not so fast

Reference 10

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:38.640017Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation a88e6c8e-8763-4e9c-8a24-555591d01e45 · outbound

This paper cites Hyper- kvasir, a comprehensive multi-class image and video dataset for gastrointestinal endoscopy.Sci- entific data, 7(1):283, 2020.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Hyper- kvasir, a comprehensive multi-class image and video dataset for gastrointestinal endoscopy.Sci- entific data, 7(1):283, 2020

Reference 11

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:38.628974Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation a3cde49e-68fc-4d6b-99e5-6f39d4083252 · outbound

This paper cites Natural images are more informative for inter- preting cnn activations than state-of-the-art syn- thetic feature visualizations.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Natural images are more informative for inter- preting cnn activations than state-of-the-art syn- thetic feature visualizations

Reference 12

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:38.616618Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation f006df30-a2b9-4e2f-9236-e104a43f976b · outbound

This paper cites Lof: identifying density-based local outliers.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Lof: identifying density-based local outliers

Reference 13

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:38.604337Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation 5673b415-1073-4820-9fb7-6aec8430394a · outbound

This paper cites Towards monosemanticity: Decomposing lan- guage models with dictionary learning.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Towards monosemanticity: Decomposing lan- guage models with dictionary learning

Reference 14

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:38.592874Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation 1653fb81-9721-4c44-8624-dbd6843a12da · outbound

This paper cites Deep learn- ing outperformed 136 of 157 dermatologists in a head-to-head dermoscopic melanoma image clas- sification task.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Deep learn- ing outperformed 136 of 157 dermatologists in a head-to-head dermoscopic melanoma image clas- sification task

Reference 15

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:38.580464Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation 89fb76ae-1f91-4d17-a394-84f0a348d3e4 · outbound

This paper cites Detecting shortcut learning for fair medical ai using shortcut testing.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Detecting shortcut learning for fair medical ai using shortcut testing

Reference 16

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:38.569522Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation 208f8805-a582-4bef-b691-11a7f16d9197 · outbound

This paper cites Dora: Exploring outlier representations in deep neural networks.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Dora: Exploring outlier representations in deep neural networks

Reference 17

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:38.559524Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation 66d3ec16-bd9c-4d59-8fca-7d91d36b8cf5 · outbound

This paper cites Labeling neural representations with inverse recognition.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Labeling neural representations with inverse recognition

Reference 18

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:38.549078Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation 505414a4-7557-4cf1-a3f4-505998f45ba0 · outbound

This paper cites Analysis of the isic image datasets: Usage, benchmarks and recommen- dations.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Analysis of the isic image datasets: Usage, benchmarks and recommen- dations

Reference 19

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:38.538255Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation fc9c2fd0-b5a5-46a1-8173-e168a4838f14 · outbound

This paper cites an unresolved cited work.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Unresolved cited work

Reference 20

Resolution: unresolved; raw_fallback, observed 2026-08-10T15:39:38.526272Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation d6075b2b-bbd1-4037-a0f0-50b8b32ff3ca · outbound

This paper cites Bcn20000: Dermoscopic lesions in the wild, 2019.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Bcn20000: Dermoscopic lesions in the wild, 2019

Reference 21

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:38.515217Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation 71eee9c5-0431-49cb-86a9-f77d4efa826c · outbound

This paper cites Concept activation regions: A generalized frame- work for concept-based explanations.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Concept activation regions: A generalized frame- work for concept-based explanations

Reference 22

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:38.503915Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation f86cf2d1-2079-4fcc-a622-e85831c9bf69 · outbound

This paper cites Visual-TCAV: Concept-based Attribution and Saliency Maps for Post-hoc Explainability in Image Classification.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Visual-TCAV: Concept-based Attribution and Saliency Maps for Post-hoc Explainability in Image Classification

Reference 23

Resolution: unresolved; no resolver link, observed 2026-08-10T15:39:37.457720Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation df754309-3a78-48dc-956f-ef785674a6cc · outbound

This paper cites Ai for radiographic covid-19 detection se- lects shortcuts over signal.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Ai for radiographic covid-19 detection se- lects shortcuts over signal

Reference 24

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:38.491904Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation 98e344b9-2a5c-4f9c-b833-b0287e283918 · outbound

This paper cites Imagenet: A large-scale hierarchical image database.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Imagenet: A large-scale hierarchical image database

Reference 25

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:38.480939Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation 150607b9-78c1-47b7-af28-7d2c08b41486 · outbound

This paper cites Predicting parameters in deep learning.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Predicting parameters in deep learning

Reference 26

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:38.470219Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation be988b00-94db-4f4d-b3d2-8eb059b3d7a1 · outbound

This paper cites An image is worth 16x16 words: Transformers for image recognition at scale.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data An image is worth 16x16 words: Transformers for image recognition at scale

Reference 27

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:38.459527Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation 3b3531f3-bba5-44e5-b7ba-36850b5e5d6d · outbound

This paper cites Understand- ing the (extra-) ordinary: Validating deep model decisions with prototypical concept-based expla- nations.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Understand- ing the (extra-) ordinary: Validating deep model decisions with prototypical concept-based expla- nations

Reference 28

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:38.448752Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation f6063e5b-b7ea-4a40-a7a0-e161d3a77187 · outbound

This paper cites From hope to safety: Unlearning bi- ases of deep models via gradient penalization in latent space.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data From hope to safety: Unlearning bi- ases of deep models via gradient penalization in latent space

Reference 29

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:38.438211Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation d60c39e8-ce4a-41bf-840b-214a1e09a591 · outbound

This paper cites Pure: Turning polysemantic neurons into pure features by identifying relevant cir- cuits.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Pure: Turning polysemantic neurons into pure features by identifying relevant cir- cuits

Reference 30

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:38.426206Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation 7ee80454-b736-40db-9318-288842d5d04c · outbound

This paper cites Mechanistic understanding and validation of large AI models with SemanticLens.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Mechanistic understanding and validation of large AI models with SemanticLens

Reference 31

Resolution: unresolved; no resolver link, observed 2026-08-10T15:39:37.490562Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation eec1e0d9-0215-4c87-bda4-2547cc3168a9 · outbound

This paper cites Toy Models of Superposition.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Toy Models of Superposition

Reference 32

Resolution: unresolved; no resolver link, observed 2026-08-10T15:39:37.494835Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 3a733a1d-2026-47db-b228-b07cd421222d · outbound

This paper cites Visualiz- ing higher-layer features of a deep network.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Visualiz- ing higher-layer features of a deep network

Reference 33

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:38.414859Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation cb4ab8d0-0d4d-4b21-9448-3d06116873ab · outbound

This paper cites Craft: Con- cept recursive activation factorization for ex- plainability.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Craft: Con- cept recursive activation factorization for ex- plainability

Reference 34

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:38.404023Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation 7a7ae970-7745-4280-bc06-00ec85bccf35 · outbound

This paper cites Unlocking feature visu- alization for deep network with magnitude con- strained optimization.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Unlocking feature visu- alization for deep network with magnitude con- strained optimization

Reference 35

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:38.393957Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation acb678db-49fb-444d-b8a4-6276f52092c7 · outbound

This paper cites A holis- tic approach to unifying automatic concept ex- traction and concept importance estimation.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data A holis- tic approach to unifying automatic concept ex- traction and concept importance estimation

Reference 36

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:38.383325Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation 30a03ec9-40ec-4df8-b212-281dbe52aae2 · outbound

This paper cites The use of multiple measure- ments in taxonomic problems.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data The use of multiple measure- ments in taxonomic problems

Reference 37

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:38.372084Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation 79349a6d-172f-4b76-bced-0b23ed4fbbaa · outbound

This paper cites Net2vec: Quan- tifying and explaining how concepts are encoded by filters in deep neural networks.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Net2vec: Quan- tifying and explaining how concepts are encoded by filters in deep neural networks

Reference 38

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:38.361541Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation 8f8565b1-c9b8-42d0-82b6-b40524b3c691 · outbound

This paper cites Shortcut learning in deep neural networks.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Shortcut learning in deep neural networks

Reference 39

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:38.349778Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation 15c4a6be-78b1-498e-9d68-4839d1eef135 · outbound

This paper cites Towards automatic concept- based explanations.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Towards automatic concept- based explanations

Reference 40

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:38.338764Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation 830a9841-c203-43b2-a27d-a456f3df9c08 · outbound

This paper cites Concept discovery and dataset exploration with singular value decomposition.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Concept discovery and dataset exploration with singular value decomposition

Reference 41

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:38.328430Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation 441245b8-8eac-407e-a7b1-3612e45b9d7f · outbound

This paper cites Deep residual learning for image recog- nition.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Deep residual learning for image recog- nition

Reference 42

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:38.317968Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation 5669e98b-5afb-4a1a-a422-b8c851d9c7e3 · outbound

This paper cites Bag of tricks for image classification with convolutional neural networks.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Bag of tricks for image classification with convolutional neural networks

Reference 43

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:38.307431Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation b65cae46-ba3d-4fa6-88ae-020a61a38cfb · outbound

This paper cites Natural language descriptions of deep visual features.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Natural language descriptions of deep visual features

Reference 44

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:38.296248Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation b9a8a6ff-e84a-425c-b639-a7f6fdd3dcb5 · outbound

This paper cites Sparse autoencoders find highly interpretable features in language models.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Sparse autoencoders find highly interpretable features in language models

Reference 45

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:38.284795Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation dec37b2e-38af-4903-b105-ec4e51406818 · outbound

This paper cites Chexpert: A large chest radio- graph dataset with uncertainty labels and expert comparison.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Chexpert: A large chest radio- graph dataset with uncertainty labels and expert comparison

Reference 46

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:38.273289Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation 3b949265-2c1d-4473-a850-6667ea89b3d6 · outbound

This paper cites Interpretability beyond feature attribu- tion: Quantitative testing with concept activa- tion vectors (tcav).

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Interpretability beyond feature attribu- tion: Quantitative testing with concept activa- tion vectors (tcav)

Reference 47

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:38.261852Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation 2cebc29d-7468-4f8f-8526-9ae8287f2750 · outbound

This paper cites Unmasking clever hans predictors and assessing what ma- chines really learn.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Unmasking clever hans predictors and assessing what ma- chines really learn

Reference 48

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:38.249360Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation 185e58ee-c376-4121-bcf7-27802dfc4c4a · outbound

This paper cites Umap: Uniform man- ifold approximation and projection.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Umap: Uniform man- ifold approximation and projection

Reference 49

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:38.237703Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation 1755349c-420b-4245-9dac-a65c13350be0 · outbound

This paper cites Evaluating the stability of semantic concept representations in cnns for robust explainability.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Evaluating the stability of semantic concept representations in cnns for robust explainability

Reference 50

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:38.224860Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation 083a0adb-5366-4fc9-ae2d-1985d89ab478 · outbound

This paper cites Visualization of neu- ral networks using saliency maps.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Visualization of neu- ral networks using saliency maps

Reference 51

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:38.212509Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation 7a1e23f8-e523-446e-af7c-d5885828ee63 · outbound

This paper cites Spurious fea- tures everywhere-large-scale detection of harm- ful spurious features in imagenet.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Spurious fea- tures everywhere-large-scale detection of harm- ful spurious features in imagenet

Reference 52

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:38.201391Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation c918585a-5499-4008-af53-919f521416b5 · outbound

This paper cites Clip- dissect: Automatic description of neuron repre- sentations in deep vision networks.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Clip- dissect: Automatic description of neuron repre- sentations in deep vision networks

Reference 53

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:38.189256Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation 34755a43-03b4-446c-b314-f7b97ec96ea8 · outbound

This paper cites Feature visualization.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Feature visualization

Reference 54

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:38.178249Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation 72c9b6de-b460-4d59-b1f1-64ae66624e20 · outbound

This paper cites Zoom in: An introduction to circuits.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Zoom in: An introduction to circuits

Reference 55

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:38.166780Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation 110ed99b-6f00-413f-bad4-3f23916e6d63 · outbound

This paper cites A threshold selection method from gray-level histograms.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data A threshold selection method from gray-level histograms

Reference 56

Resolution: unresolved; no resolver link, observed 2026-08-10T15:39:37.585767Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 39aa8105-3119-46bd-ad3c-fd8e7235d939 · outbound

This paper cites Reveal to re- vise: An explainable ai life cycle for iterative bias correction of deep models.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Reveal to re- vise: An explainable ai life cycle for iterative bias correction of deep models

Reference 57

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:38.146555Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation 2b6ec17c-b697-46a4-a744-63e51949c802 · outbound

This paper cites Navigating neural space: Revisiting concept activation vectors to over- come directional divergence.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Navigating neural space: Revisiting concept activation vectors to over- come directional divergence

Reference 58

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:38.134907Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation 35c8e584-c8ef-4736-bcfc-ab4da79066ae · outbound

This paper cites Py- torch: An imperative style, high-performance deep learning library.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Py- torch: An imperative style, high-performance deep learning library

Reference 59

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:38.122810Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation e07d5fd4-eafc-4dfd-96b0-e091a48eb28a · outbound

This paper cites Interpretable data-based expla- nations for fairness debugging.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Interpretable data-based expla- nations for fairness debugging

Reference 60

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:38.110974Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation 32f344b7-09d9-454b-8a31-1246f701f416 · outbound

This paper cites Learning to Generate Reviews and Discovering Sentiment.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Learning to Generate Reviews and Discovering Sentiment

Reference 61

Resolution: unresolved; no resolver link, observed 2026-08-10T15:39:37.606346Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation c6aa22e3-90d8-45d3-88a2-61c676666b44 · outbound

This paper cites Interpretations are useful: penalizing explanations to align neural networks with prior knowledge.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Interpretations are useful: penalizing explanations to align neural networks with prior knowledge

Reference 62

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:38.098189Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation f3300fa2-c7f5-4d2e-9463-8162ef0e97b0 · outbound

This paper cites Right for the right reasons: training differentiable models by constraining their explanations.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Right for the right reasons: training differentiable models by constraining their explanations

Reference 63

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:38.085617Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation 0c580185-4688-4bd4-83a2-6ef7303401a9 · outbound

This paper cites Making deep neural networks right for the right scientific rea- sons by interacting with their explanations.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Making deep neural networks right for the right scientific rea- sons by interacting with their explanations

Reference 64

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:38.074980Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation fb723d17-98a7-42f0-99a7-406f22b85b0e · outbound

This paper cites Grad-cam: Visual explanations from deep networks via gradient- based localization.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Grad-cam: Visual explanations from deep networks via gradient- based localization

Reference 65

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:38.062752Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation c2474dd2-5796-4a33-8fd4-6eeaa2f1d025 · outbound

This paper cites Very deep convolutional networks for large-scale im- age recognition.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Very deep convolutional networks for large-scale im- age recognition

Reference 66

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:38.052012Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation d9d7d3a4-21b2-46b2-ad86-f65c7d817288 · outbound

This paper cites Salient imagenet: How to discover spurious features in deep learning.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Salient imagenet: How to discover spurious features in deep learning

Reference 67

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:38.040180Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation 6b5f584a-3e59-4182-9e61-ae01edca9867 · outbound

This paper cites Explaining ma- chine learning models for clinical gait analysis.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Explaining ma- chine learning models for clinical gait analysis

Reference 68

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:38.027233Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation 9407f59a-33e1-43b2-bcff-c0b489bb16f0 · outbound

This paper cites Deep learning for ecg analysis: Benchmarks and insights from ptb- xl.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Deep learning for ecg analysis: Benchmarks and insights from ptb- xl

Reference 69

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:38.015817Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation d5f57d7d-33e3-4843-945b-3a01f3bfa1f6 · outbound

This paper cites Intriguing prop- erties of neural networks.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Intriguing prop- erties of neural networks

Reference 70

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:38.003580Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation 87271835-a6d9-4043-a6a8-baaf8b3b3a4b · outbound

This paper cites The ham10000 dataset, a large collection of multi-source dermatoscopic images of com- mon pigmented skin lesions.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data The ham10000 dataset, a large collection of multi-source dermatoscopic images of com- mon pigmented skin lesions

Reference 71

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:37.992309Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation 5eb92ef8-b4ef-4824-95b6-c114315ef0ae · outbound

This paper cites Visualizing data using t-sne.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Visualizing data using t-sne

Reference 72

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:37.981146Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation bf5a2673-44e4-4789-9cba-b61dfbd5d6cd · outbound

This paper cites Multi-dimensional concept discovery (mcd): A unifying framework with completeness guarantees.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Multi-dimensional concept discovery (mcd): A unifying framework with completeness guarantees

Reference 73

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:37.968769Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation 1546d967-2d72-4bf1-a405-d7583965cc5e · outbound

This paper cites Ptb- xl, a large publicly available electrocardiography dataset.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Ptb- xl, a large publicly available electrocardiography dataset

Reference 74

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:37.958072Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation 8a108c6f-d902-4e66-98bd-d160eb7f3de3 · outbound

This paper cites Explaining deep learning for ecg analysis: Building blocks for auditing and knowledge discovery.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Explaining deep learning for ecg analysis: Building blocks for auditing and knowledge discovery

Reference 75

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:37.945505Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation 8b3c8565-4d0e-41c4-a059-cda7d36cad82 · outbound

This paper cites Fast diffusion-based counterfactuals for shortcut re- moval and generation.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Fast diffusion-based counterfactuals for shortcut re- moval and generation

Reference 76

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:37.933134Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation 7c8e927a-d8fe-4306-bfe8-90fe88e6dcb3 · outbound

This paper cites Pytorch image mod- els.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Pytorch image mod- els

Reference 77

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:37.922681Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation 6bae9e2a-1260-411a-8436-2bf589e51d84 · outbound

This paper cites Discover and cure: Concept- aware mitigation of spurious correlation.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Discover and cure: Concept- aware mitigation of spurious correlation

Reference 78

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:37.911965Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation d5685515-2862-4741-8bd3-2198d78d0144 · outbound

This paper cites Variable generalization performance of a deep learning model to de- tect pneumonia in chest radiographs: a cross- sectional study.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Variable generalization performance of a deep learning model to de- tect pneumonia in chest radiographs: a cross- sectional study

Reference 79

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:37.901306Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation c09e634a-57e6-4190-afc3-1ad8f9149c09 · outbound

This paper cites Invertible concept-based explanations for cnn models with non-negative concept activa- tion vectors.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Invertible concept-based explanations for cnn models with non-negative concept activa- tion vectors

Reference 80

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:37.888808Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation 380f1e53-ccc0-4f04-b1f7-cb5ee53ef014 · outbound

This paper cites right-reason.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data right-reason

Reference 81

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:37.876041Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation 8895df10-f0fd-46d0-986f-0071a0725892 · outbound

This paper cites Slic superpixels compared to state- of-the-art superpixel methods.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Slic superpixels compared to state- of-the-art superpixel methods

Reference 82

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:37.864227Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation 9ee809f4-98b7-4a47-879c-b1891820aca9 · outbound

This paper cites Support- vector networks.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Support- vector networks

Reference 83

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:37.852956Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation 8326d7e6-d6d4-477f-8571-78c2f99dc920 · outbound

This paper cites A uni- fied approach to interpreting model predictions.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data A uni- fied approach to interpreting model predictions

Reference 84

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:37.841067Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation 352af60c-2dce-4bd5-88f2-1d95a322e05a · outbound

This paper cites Beyond word importance: Contextual decompo- sition to extract interactions from lstms.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Beyond word importance: Contextual decompo- sition to extract interactions from lstms

Reference 85

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:37.828873Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation 3f02e7b2-7b89-4911-b069-0a8bda0a1f80 · outbound

This paper cites Null it out: Guarding protected attributes by iterative nullspace projection.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Null it out: Guarding protected attributes by iterative nullspace projection

Reference 86

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:37.817340Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Observation d100b99e-dbeb-43b5-b313-33c407701189 · outbound

This paper cites Editing a classifier by rewriting its prediction rules.

Ensuring Medical AI Safety: Interpretability-Driven Detection and Mitigation of Spurious Model Behavior and Associated Data Editing a classifier by rewriting its prediction rules

Reference 87

Resolution: verified fuzzy; raw_fallback, observed 2026-08-10T15:39:37.804317Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-20T06:33:59.587034+00:00.

Coverage vector

Reference resolution

Outbound references

Pith citing papers