Latest Interpretability Research Papers
The newest Interpretability papers from across the field — arXiv, NeurIPS, CVPR, Nature, and more — refreshed daily and ranked by relevance. Distill AI tracks Interpretability so you don’t have to: get the standout work delivered to your inbox every morning, with 2-sentence summaries and the option to chat with any paper.
Get the latest Interpretability papers in your inbox — free →Recent papers
- Post-Interpretive Criticism: Volume III - The Canon of WitnessesVale, Dorian, Museum of One · Zenodo (CERN European Organ... · Apr 28, 2028
Title: The Canon of Witnesses Type: Publication > Journal Volume Journal Title: The Journal of Post-Interpretive Criticism Volume: 3 (upcoming) ISSN: 2819-7232 Author: Dorian Vale Publisher: Museum of One DOI: 10.5281/zenodo.17421408 Museum…
- Post-Interpretive Criticism: Volume III - The Canon of WitnessesVale, Dorian, Museum of One · Zenodo (CERN European Organ... · Apr 28, 2028
Title: The Canon of Witnesses Type: Publication > Journal Volume Journal Title: The Journal of Post-Interpretive Criticism Volume: 3 (upcoming) ISSN: 2819-7232 Author: Dorian Vale Publisher: Museum of One DOI: 10.5281/zenodo.17421408 Museum…
- Post-Interpretive Criticism: Volume III - The Canon of WitnessesVale, Dorian, Museum of One · Zenodo (CERN European Organ... · Apr 28, 2028
Title: The Canon of Witnesses Type: Publication > Journal Volume Journal Title: The Journal of Post-Interpretive Criticism Volume: 3 (upcoming) ISSN: 2819-7232 Author: Dorian Vale Publisher: Museum of One DOI: 10.5281/zenodo.17421408 Museum…
- Development and validation of a machine learning-based predictive model for prognosis in cerebral hemorrhage patients after hyperbaric oxygen therapy.Congyi Ding, Jun Wu, Guangning Zhang, Xiaoming Feng et al. · PubMed · Jan 1, 2027
Intracerebral hemorrhage is a common and critical neurological emergency, and the inconsistent results of hyperbaric oxygen therapy make predicting outcomes challenging. Therefore, a reliable predictive model for outcomes is needed to impro…
- Deciphering the threat: informational and analytical foundations for police training on encrypted criminal networks. The cases of Ukraine, England, and GermanyBohdan-Petro Koshovyi, Andrii Manko, Pavlo P. Latkovskyi, Aleksey Demchenko · Zenodo (CERN European Organ... · Jan 1, 2027
This article interprets, from a hermeneutic and documentary perspective, the informational and analytical foundations that recent academic literature identifies as necessary for the training of police forces in dealing with encrypted crimin…
- The impact of war on the mental health of vulnerable populations in Ukraine: the role of international assistance in resilience and post-traumatic recoveryВалентина БОДАК, Юлія НІКОЛЬЧУК, Карпінська Роксолана Іванівна, Volodymyr Zhyhaylo · Zenodo (CERN European Organ... · Jan 1, 2027
This article uses a phenomenological-hermeneutic approach to explore the meaning that Ukrainian individuals and communities affected by the war attribute to psychological suffering and its relationship to international aid, moving beyond an…
- world2vec: Mechanistic world models through latent dynamics learningTobias Schmidt, Rodrigo Gonzalez Laiz, Krisztián Szuppinger, Steffen Schneider · arXiv · Oct 29, 2026
World models can achieve strong control performance without faithfully capturing the dynamics of their environment, limiting their potential for mechanistic interpretability and scientific discovery. Here, we introduce world2vec to learn me…
- Homebrewed SLMs: Autoregressive and Masked Diffusion Language Models Built at Home for PortuguesePedro Carvalho Brom · arXiv · Oct 29, 2026
Building a language model of one's own no longer requires a data centre. This tutorial trains one from scratch, tokenizer included, on one consumer GPU, and shows how to tell whether it is good for the tasks of a specialist domain. It propo…
- Discovering Interpretable Acoustic Features in Autoregressive Voice-Cloning TTS ModelsAnuj Diwan, David Harwath, Eunsol Choi · arXiv · Oct 29, 2026
Voice-cloning TTS models synthesize an unseen speaker's voice from brief speech prompts, but it remains unclear which acoustic properties they represent and use. We introduce a feature-based interpretability framework for autoregressive voi…
- Multistep and Multivariate Time Series Explainability with GradientSHAP: A Wind Power Case StudyRobbie Southam, Zhong Fan, Ke Li · arXiv · Oct 29, 2026
As the effects of climate change become increasingly apparent, the demand for renewable energy sources such as wind power continues to grow. However, renewables' variable and intermittent nature presents significant challenges in their inte…
- Exploring the Potential of Symbolic Regression in Precision Livestock FarmingElisabeth Mayrhuber, Maciej Oczak, Michael Iwersen, David Brunner et al. · arXiv · Oct 29, 2026
The ability to predict farrowing time in pigs is crucial for optimal farm management and ensuring piglet survival. This study investigates the potential of explainable artificial intelligence for precision livestock farming, focusing on acc…
- Multi-output LSTM soft sensors for industrial anaerobic co-digestion using LSTM-beta-VAE generative data augmentationRohit Murali, Benaissa Dekhici, Tao Chen, Dongda Zhang et al. · arXiv · Oct 29, 2026
Anaerobic digestion (AD) requires close monitoring of key process variables, including pH, the ratio of volatile organic acids to alkaline buffer capacity and biogas production, to maintain biological stability and prevent process failure. …
- Improving the Intrinsic Interpretability of Differentiable Decision Trees through Euclidean Feature-Concentration RegularisationQiqi Yang · arXiv · Oct 29, 2026
Differentiable decision trees (DDTs) support end-to-end optimisation and explicit decision paths, but dense oblique splits can obscure the role of individual input features. We propose the Euclidean-Concentrated Differentiable Decision Tree…
- Explainable and Adaptive Machine Learning for Fraud Risk Detection in Bangladesh's Digital Banking EcosystemMd Najmus Swaqeeb · arXiv · Oct 29, 2026
The rapid expansion of digital banking and mobile financial services creates growing challenges for fraud detection, particularly under severe class imbalance and operational constraints. This study presents an explainable and adaptive mach…
- Interpretable by Design Tabular Prediction via In-Context Local Linear LearningRan Eisenberg, Steven Berger, Ofir Lindenbaum · arXiv · Oct 29, 2026
In-context learning (ICL) lets tabular foundation models make predictions without dataset-specific training or fine-tuning. These models achieve strong predictive accuracy, but interpreting how individual features contribute to the predicti…
- Does Federated Training Degrade Explainability? A Comparative Analysis of Grad-CAM Faithfulness in Centralized vs. Federated Deepfake DetectionMuhammad Uzair Shahid, Muhammad Zabil Mehboob · arXiv · Oct 29, 2026
Automated detection has become the essential safeguard against realistic synthetic media, but deploying it at scale requires training on sensitive data from multiple organizations, putting privacy and the transparency demanded by forensic r…
- Sparse Autoencoders as Implicit Diversification Regularisers for Cross-Sectional Cryptocurrency ReturnsWeijia Han, Lisha Qu · arXiv · Oct 29, 2026
Factor models are central to asset pricing and portfolio construction, but neural factors are often difficult to interpret, and their effect on portfolio risk is unclear. A sparse autoencoder (SAE) may address both concerns by describing da…
- Closed Gates Leak: Decomposing Adjacent-Layer Cancellation in Gated Language ModelsHarvey Dam, Ganesh Gopalakrishnan · arXiv · Oct 28, 2026
In transformer language models with gated multilayer perceptrons (MLPs), adjacent layers write partly against each other into the residual stream. Previous work interpreted this cancellation as layers proposing features and the next layer s…
- LCM-Promo: An LLM-Based Uplift Model for Promotion AllocationNesta Midavaine, Floris Fok, Paulo Mendes, Joao Silva et al. · ECML/PKDD 2026 · Oct 27, 2026
We present LCM-Promo, an LLM-based uplift model for promotion allocation. LCM-Promo is a task-specific Large Commerce Model (LCM): a finetuned LLM trained on transactional commerce data to predict user intent and promotion response. It lear…
- Orthogonal and Parametrized Multi-Attribute Steering of LLMsRemy Ogasawara, Cullen Anderson, Narmeen Fatimah Oozeer, Foad Namjoo et al. · arXiv · Oct 27, 2026
Activation steering offers an effective and efficient way to modify an LLM's behavior at inference time. However, it remains unclear how to apply steering such that the signal is consistently strong and the steering strength has a universal…
- Interpretable Meta-Policies for Neural Decision MakingRaban Emunds, Jannis Blüml, Quentin Delfosse, Kristian Kersting · arXiv · Oct 27, 2026
Deep Reinforcement Learning remains difficult to interpret and align due to the black-box nature of neural networks. While (neuro-)symbolic approaches improve transparency by incorporating symbolic reasoning, applying them directly to low-l…
- Mechanistic Failure Modes of Weight Pruning in Transformer Language ModelsNabeelah Faumi, Deshan Sumanathilaka · arXiv · Oct 27, 2026
Pruned language models are judged almost entirely by accuracy and perplexity, which assumes that preserved output behaviour implies preserved internal computation. We test this assumption with a factorial pruning study on GPT-2, crossing tw…
- PREEN: Exemplar-Based Color Style Transfer with Smartphone Editor PresetsBeomjin Ahn · arXiv · Oct 27, 2026
We present PREEN (Preset Regression via an Exemplar-based Editor Network), which casts color style transfer as regression onto the 29 parameters of a deterministic image editor. Given a content image and a style reference, PREEN predicts a …
- Deep Bayesian Approaches for Interpretable Nonlinear Structural System IdentificationWei Liu, Konstantinos Vlachas, Eleni Chatzi · arXiv · Oct 27, 2026
System identification, namely the inference of mathematical representations of dynamical systems from measured data, is fundamental to engineering diagnostics, prognostics, and decision-making. Bayesian filtering offers a principled framewo…
- Structure, Not Sequence: Tracking and Validating Folding Mechanisms in ESMFold's Trunk with Sparse AutoencodersBohan Liu, Xiaohan Fan, Wenqian Ye, Aidong Zhang · arXiv · Oct 26, 2026
Structure predictors like ESMFold fold proteins using neural network computation and have transformed structural biology. However, the driving mechanism from amino-acid (AA) sequence to 3D structure is insufficiently understood, and biologi…
- CounterFlow: Target-Conditioned Flow Matching for Visual CounterfactualsPaul Sevellec, Elisa Fromont, Romaric Gaudel, Laurence Roze et al. · arXiv · Oct 26, 2026
In eXplainable AI, counterfactual explanations have gained significant traction by offering users direct, intuitive, and actionable insights into model decisions. We introduce CounterFlow, a novel framework based on Conditional Flow Matchin…
- A Hybrid Framework for Explainable Aviation Risk: Causal Chain Extraction from Safety Reports with Physics-Based Failure Quantification and Interventional AnalysisYutian Pang, Yifang Tian, Jiacheng Xie, Xinyu Zhao et al. · arXiv · Oct 26, 2026
Aviation authorities hold decades of accident reports whose narratives record how investigators reconstructed each accident. Computational work on these archives has mostly sorted reports into fixed taxonomies. In this work, we propose a hy…
- Active Learning of Conditional Generative Models via the Transport Neural Tangent KernelJayoung Ryu, Kyunghyun Cho, Romain Lopez · arXiv · Oct 26, 2026
Many high-throughput scientific experiments produce distributional outputs, requiring conditional generative models to interpret the data, and to predict outcomes at new inputs. Acquiring experimental data at new conditions (e.g., for a new…
- Adversary Brains: Hypothesis-Based Logical Signatures for Threat Actor Detection with LLMsGadi Evron · arXiv · Oct 26, 2026
An adversary brain is an LLM-based model of a threat actor's behavior, held as an evolving logical signature: a hypothesis with enough structure that data can test it continuously, without a person re-interpreting it each time. Rather than …
- Distilling neural networks into equations to predict planetary instabilitySimon Alford, Edward Gu, Omar Costilla-Reyes, Dan Tamayo et al. · arXiv · Oct 26, 2026
The problem of predicting when a multi-planet system will become dynamically unstable has defied analytical progress over the past three hundred years. Recent machine learning advances achieve high predictive performance of instability dyna…