arXiv Machine Learning
Jun 29
The Curse of Multiple Mediators: Hidden Interaction Effects in Activation Patching
arXiv:2606. 27510v1 Announce Type: new Abstract: Activation patching is the primary tool in mechanistic interpretability.
The Representational Limit of Scalar Interactions: An Interventional Decomposition
arXiv:2606. 19410v1 Announce Type: cross Abstract: Signed pairwise interaction scores fundamentally conflate uniqueness (U), redundancy (R), and synergy (S).
Related stories
arXiv Machine Learning
Jul 10
Certified Interventional Fidelity: Anytime-Valid, Adaptive Evaluation of Causal Claims in Mechanistic Interpretability
arXiv:2607. 08349v1 Announce Type: new Abstract: Mechanistic interpretability often evaluates explanations by intervening on a model: swapping hidden states, patching activations, ablating components, or comparing a compressed model to the original one.
arXiv Machine Learning
Aug 3
Nonparametric Partial Disentanglement via Mechanism Sparsity: Sparse Actions, Interventions and Sparse Temporal Dependencies
arXiv:2401. 04890v2 Announce Type: replace-cross Abstract: This work introduces a novel principle for disentanglement we call mechanism sparsity regularization, which applies when the latent factors of interest depend sparsely on observed auxiliary variables and/or past latent factors.
arXiv Machine Learning
Jul 1
When few labeled target data suffice: a theory of semi-supervised domain adaptation via fine-tuning from multiple adaptive starts
arXiv:2507. 14661v2 Announce Type: replace-cross Abstract: Semi-supervised domain adaptation (SSDA) seeks to achieve accurate predictions in a target domain with limited labeled target data by exploiting abundant source and unlabeled target data.
arXiv AI
Jun 19
Latent Confounded Causal Discovery via Lie Bracket Geometry
arXiv:2606. 19610v1 Announce Type: cross Abstract: Recent work on Kan-Do-Calculus (KDC) has established that the boundary between passive observation and active intervention in causal inference is a category-theoretic bi-adjunction, with interventions modeled by left Kan extensions and conditioning by right Kan extensions.
arXiv Machine Learning
Jul 7
Significance-First Splitting: Aligning Treatment Heterogeneity Detection with Honest Estimation
arXiv:2607. 03999v1 Announce Type: cross Abstract: Estimating heterogeneous treatment effects (CATE) requires simultaneously detecting effect modification and quantifying estimation uncertainty.
arXiv AI
Jun 30
Feature-level Interaction Explanations in Multimodal Transformers
arXiv:2603. 13326v2 Announce Type: replace-cross Abstract: Multimodal Transformers often produce predictions without clarifying how different modalities jointly support a decision.
arXiv AI
Jun 17
FoundCause: Causal Discovery with Latent Confounders from Observational Data
arXiv:2606. 17516v1 Announce Type: cross Abstract: Causal discovery from observational data remains challenging due to the need to recover directed structure and latent confounding without interventions.
arXiv Machine Learning
Sep 1
Jigsaw-CRL: Recovering Global Latent Causal Order from Fragmented Multi-Client Interventions
arXiv:2608.28991v1 Announce Type: cross Abstract: Causal representation learning (CRL) aims to recover latent causal variables and their structural relations from high-dimensional observations. Exist...
arXiv AI
4d ago
Representational Simplicity and Circuit Size Dissociate in a Threshold-Dependent Way: A Controlled Test via Adversarial Training
arXiv:2609.35890v1 Announce Type: new Abstract: Sparse-autoencoder decomposability and concentrated feature attribution are increasingly treated as evidence that a model's computation is easier to re...
arXiv Machine Learning
Aug 31
Actionable CBFI: Integrating Structural Decomposition and Causal Counterfactual Recourse for Tabular Machine Learning
The paper introduces Actionable Case-Based Feature Importance (A‑CBFI), a framework that integrates structural causal models with counterfactual recourse for tabular machine learning. A‑CBFI isolates synergistic interaction bottlenecks and releases suppressive structural locks, concentrating over 98.3% of intervention effort on diagnosed root causes. Empirical tests in finance and healthcare show a 76.9% reduction in active human intervention while keeping recourse costs comparable to exhaustive causal methods.
arXiv Machine Learning
Jul 10
Partial Causal Structure Learning for Valid Selective Conformal Inference under Interventions
arXiv:2603. 02204v2 Announce Type: replace Abstract: Selective conformal prediction can yield substantially tighter uncertainty sets when we can identify calibration examples that are exchangeable with the test example.