Causal Scrubbing

Evaluation considerations

Two popular approaches to validate causality are causal scrubbing and interchange interventions (Chan et al., 2022a; Geiger et al., 2024).

Together, these methods provide stronger evidence for mechanistic explanations than observational analysis alone.

Intended applications

Circuit faithfulness: demonstrating circuits are necessary and sufficient for explaining behaviour via causal intervention (e.g., causal scrubbing (Chan et al., 2022a)), without performance degradation.

Limitations

By aggregating faithfulness scores across tasks and settings, benchmarks such as MIB (Mueller et al., 2025) evaluate how well circuit discovery methods identify causally important model components. However, achieving strict necessity and sufficiency is often hard because many model behaviours depend on distributed representations and multiple redundant pathways (Barez et al., 2026).

Even if we identify a circuit that captures the primary mechanism of a behaviour, components outside the circuit may still play important roles, such as handling edge cases and computational robustness.

Related methods

• Interchange Interventions

Compare side by side →

Key papers

Cited in §7.3.

• ■ Michael Hanna, Sandro Pezzelle, and Yonatan Belinkov. Have faith in faithfulness: Going beyond circuit overlap when finding model mechanisms. In First Conference on Language Modeling, 2024. URL https://openreview.net/forum?id=TZ0CCGDcuT.
• ■ Samuel Marks, Can Rager, Eric J Michaud, Yonatan Belinkov, David Bau, and Aaron Mueller. Sparse feature circuits: Discovering and editing interpretable causal graphs in language models. In The Thirteenth International Conference on Learning Representations, 2025a. URL https://openreview.net/forum?id=I4e82CIDxv.
• ■ PreprintLawrence Chan, Adrià Garriga-Alonso, Nicholas Goldwosky-Dill, Ryan Greenblatt, Jenny Nitishinskaya, Ansh Radhakrishnan, Buck Shlegeris, and Nate Thomas. Causal scrubbing, a method for rigorously testing interpretability hypotheses. AI Alignment Forum, 2022a. https://www.alignmentforum.org/posts/JvZhhzycHu2Yd57RN/causal-scrubbing-a-method-for-rigorously-testing.
• ■ Atticus Geiger, Zhengxuan Wu, Christopher Potts, Thomas Icard, and Noah Goodman. Finding alignments between interpretable causal variables and distributed neural representations. In Francesco Locatello and Vanessa Didelez (eds.), Proceedings of the Third Conference on Causal Learning and Reasoning, volume 236 of Proceedings of Machine Learning Research, pp. 160–187. PMLR, 01–03 Apr 2024. URL https://proceedings.mlr.press/v236/geiger24a.html.
• ■ Tung-Yu Wu and Fazl Barez. Query circuits: Explaining how language models answer user prompts. In Forty-third International Conference on Machine Learning, 2026. URL https://openreview.net/forum?id=7F0sragazb.
• ■ Aaron Mueller, Atticus Geiger, Sarah Wiegreffe, Dana Arad, Iván Arcuschin, Adam Belfki, Yik Siu Chan, Jaden Fried Fiotto-Kaufman, Tal Haklay, Michael Hanna, Jing Huang, Rohan Gupta, Yaniv Nikankin, Hadas Orgad, Nikhil Prakash, Anja Reusch, Aruna Sankaranarayanan, Shun Shao, Alessandro Stolfo, Martin Tutek, Amir Zur, David Bau, and Yonatan Belinkov. MIB: A mechanistic interpretability benchmark. In Forty-second International Conference on Machine Learning, 2025. URL https://openreview.net/forum?id=sSrOwve6vb.
• ■ Fazl Barez, Tung-Yu Wu, Iván Arcuschin, Michael Lan, Vincent Wang, Noah Siegel, Nicolas Collignon, Clement Neo, Isabelle Lee, Alasdair Paren, et al. Chain-of-thought is not explainability. In The Fortieth Annual Conference on Neural Information Processing Systems Position Paper Track, 2026.

Other methods in Circuits in Mechanistic Interpretability

Activation Patching · Path Patching · ACDC · Circuit Tracing (SAE- or CLT-Based Circuit Discovery) · Interchange Interventions