cs.LGNov 16, 2026

Programmable Semantic Interventions: Learning Causal Carriers from Behavioral Programs

Authors: Paolo Muratore, Mackenzie Weygandt Mathis

Organizations: EPFL

Abstract

Causal abstraction explains a frozen network by aligning its activations with the variables of a simpler, interpretable model, and checking that interventions on the two agree. Finding such an alignment currently requires a counterfactual dataset: for every pair of prompts, the answer the network ought to give once the variable has been changed. In practice, those labels are produced by an executable algorithm for the behavior, written before studying the model supposed to implement it. However, the criterion these methods optimize only leverages the answer an intervention should produce. Here, we develop an approach that lets the investigator define that relationship directly: the answers the variable can produce, and which of them an intervention should yield. We call this interface a semantic program and introduce Programmable Semantic Interventions (PSI): a low-rank activation subspace learned so that interchanging it changes the model's answer as the program declares, with the expected answers read from the frozen model itself. On the MIB benchmark PSI performs on par with supervised alignment at matched data and training budget, without curated pairs, and it needs no ground-truth labels; while on natural science questions with gold labels withheld, it tracks the model's own answers better than a subspace trained on gold labels. The declaration also fixes a coordinate system with known semantics, enabling donor-free steering: jointly training PSI for both localization and control realizes 94.8% of held-out requests for answers the investigator chooses, without the need for donor prompts.

Figures & tables

Appendix figures & tables11 assets

Supplementary material from the paper’s appendix.

Appendix

Explore similar work

CardsList
  1. Certifying Interventional Agreement Among Observationally Equivalent Causal Models

    Sep 26, 2026Sourena Khanzadeh, Daniel Platnick, Marjan Alirezaie +1Causal InferenceCausal Intervention

  2. Intervene, Calibrate, Discover: Learning Causal Structure from Pixels

    Oct 26, 2026Ziyi Ding, Xiaodi Huang, Xianxin Lai +4Causal Representation LearningCausal Discovery

  3. Predictive gains, causal losses: why interventional competence cannot be learned from data alone

    Nov 2, 2026Marcus ThomasCausal InterventionCausal Inference