mlsanityData to 7 Oct 2026

Causal Abstraction with Soft Interventions

arXiv 2211.12270 ↗First posted 22 Nov 2022cs.AIcited by 23

Not on the map. Why?

Abstract (arXiv)

Causal abstraction provides a theory describing how several causal models can represent the same system at different levels of detail. Existing theoretical proposals limit the analysis of abstract models to "hard" interventions fixing causal variables to be constant values. In this work, we extend causal abstraction to "soft" interventions, which assign possibly non-constant functions to variables without adding new causal connections. Specifically, (i) we generalize τ-abstraction from Beckers and Halpern (2019) to soft interventions, (ii) we propose a further definition of soft abstraction to ensure a unique map ω between soft interventions, and (iii) we prove that our constructive definition of soft abstraction guarantees the intervention map ω has a specific and necessary explicit form.

Key facts. Cited by 23 papers in the map (data through 7 Oct 2026). Builds on 10 of its 24 references that are in the map. First posted on arXiv 22 Nov 2022.

Builds on

10 of 24 references.

References · 10

  1. 2021Causal Abstractions of Neural Networkscited by 401
  2. 2016Foundations of Structural Causal Models with Cycles and Latent Variablescited by 134
  3. 2017Causal Consistency of Structural Equation Modelscited by 114
  4. 2018Abstracting Causal Modelscited by 104
  5. 2019Approximate Causal Abstractioncited by 62
  6. 2014Visual Causal Feature Learningcited by 59
  7. 2022Causal Structure Learning: a Combinatorial Perspectivecited by 46
  8. 2021Compositional Abstraction Error and a Category of Causal Modelscited by 32
  9. 2016Unsupervised Discovery of El Nino Using Causal Feature Learning on Microlevel Climate Datacited by 19
  10. 2022Abstraction between Structural Causal Models: A Review of Definitions and Propertiescited by 14

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2023: 6; 2024: 3; 2025: 6; 2026: 8.

23 papers.

Sort:

  1. 2023Finding Alignments Between Interpretable Causal Variables and Distributed Neural Representationscited by 197
  2. 2023Causal Abstraction: A Theoretical Foundation for Mechanistic Interpretabilitycited by 194
  3. 2023The Hydra Effect: Emergent Self-repair in Language Model Computationscited by 145
  4. 2024The Landscape of Causal Discovery Data: Grounding Causal Discovery in Real-World Applicationscited by 17
  5. 2023Targeted Reduction of Causal Modelscited by 15
  6. 2023Causal Optimal Transport of Abstractionscited by 13
  7. 2025I Predict Therefore I Am: Is Next Token Prediction Enough to Learn Human-Interpretable Concepts from Data?cited by 13
  8. 2024Aligning Graphical and Functional Causal Abstractionscited by 6
  9. 2025What is causal about causal models and representations?cited by 6
  10. 2023Meaningful Causal Aggregation and Paradoxical Confoundingcited by 5
  11. 2025Combining Causal Models for More Accurate Abstractions of Neural Networkscited by 4
  12. 2024Causal Abstraction in Model Interpretability: A Compact Surveycited by 2
11 more
  1. 2026Certified Interventional Fidelity: Anytime-Valid, Adaptive Evaluation of Causal Claims in Mechanistic Interpretabilitycited by 2
  2. 2026Hidden APIs in Language Models: Discovering Reusable Causal Interfaces from Forked Futurescited by 2
  3. 2025On the Identifiability of Causal Abstractionscited by 1
  4. 2025Learning Nonlinear Causal Reductions to Explain Reinforcement Learning Policiescited by 1
  5. 2026Causal Mechanism Reduction: Mechanism Replacement for Neural Network Pruning and Abstractioncited by 1
  6. 2025Cyclic Counterfactuals under Shift-Scale Interventionscited by 0
  7. 2026Concept Component Analysis: A Principled Approach for Concept Extraction in LLMscited by 0
  8. 2026Causal Foundations of Collective Agencycited by 0
  9. 2026Operationalising Relative Causal Knowledge: Backbone Identifiability from Private Reports on a Shared Outcomecited by 0
  10. 2026Local Gains and Fixed-Assignment Set Losses in Shared Set Decoderscited by 0
  11. 2026Score-Based Learning of Cluster DAGs from Interventionscited by 0