Towards Automated Circuit Discovery for Mechanistic Interpretability
Open accessThe DOI matches its OpenAlex record; 15 of the 15 references it cites were found in a registry or an index.
Abstract
Through considerable effort and intuition, several recent works have reverse-engineered nontrivial behaviors of transformer models. This paper systematizes the mechanistic interpretability process they followed. First, researchers choose a metric and dataset that elicit the desired model behavior. Then, they apply activation patching to find which abstract neural network units are involved in the behavior. By varying the dataset, metric, and units under investigation, researchers can understand the functionality of each component. We automate one of the process' steps: to identify the circuit that implements the specified behavior in the model's computational graph. We propose several algorithms and reproduce previous interpretability results to validate them. For example, the ACDC algorithm rediscovered 5/5 of the component types in a circuit in GPT-2 Small that computes the Greater-Than operation. ACDC selected 68 of the 32,000 edges in GPT-2 Small, all of which were manually found by previous work. Our code is available at https://github.com/ArthurConmy/Automatic-Circuit-Discovery.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Categorical Reparameterization with Gumbel-Softmax
Jang, E., Gu, S., & Poole, B. (2016) · arXiv · doi:10.48550/arxiv.1611.01144 · Cited by 3,616
Emergent Abilities of Large Language Models
Wei, J., Tay, Y., Bommasani, R., Raffel, C., Zoph, B., Borgeaud, S., Yogatama, D., Bosma, M., Zhou, D., Metzler, D., et al. (2022) · arXiv · doi:10.48550/arxiv.2206.07682 · Cited by 1,264
Interpretable Explanations of Black Boxes by Meaningful Perturbation
Fong, R. C., & Vedaldi, A. (2017) · 2017 IEEE International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2017.371 · Cited by 1,097
A Survey on Neural Network Interpretability
Zhang, Y., Tino, P., Leonardis, A., & Tang, K. (2021) · IEEE Transactions on Emerging Topics in Computational Intelligence · doi:10.1109/tetci.2021.3100641 · Cited by 923
Olah, C., Mordvintsev, A., & Schubert, L. (2017) · Distill · doi:10.23915/distill.00007 · Cited by 880
The Malicious Use of Artificial Intelligence: Forecasting, Prevention, and Mitigation
Brundage, M., Avin, S., Clark, J., Toner, H., Eckersley, P., Garfinkel, B., Dafoe, A., Scharre, P., Zeitzoff, T., Filar, B., et al. (2018) · Apollo - University of Cambridge Repository · doi:10.17863/cam.22520 · Cited by 849
Towards Faithfully Interpretable NLP Systems: How Should We Define and Evaluate Faithfulness?
Jacovi, A., & Goldberg, Y. (2020) · Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics · doi:10.18653/v1/2020.acl-main.386 · Cited by 503
Analysis of Explainers of Black Box Deep Neural Networks for Computer Vision: A Survey
Buhrmester, V., Münch, D., & Arens, M. (2021) · Machine Learning and Knowledge Extraction · doi:10.3390/make3040048 · Cited by 403
Transformer Feed-Forward Layers Are Key-Value Memories
Geva, M., Schuster, R., Berant, J., & Levy, O. (2020) · arXiv · doi:10.48550/arxiv.2012.14913 · Cited by 353
Transformer Feed-Forward Layers Are Key-Value Memories
Geva, M., Schuster, R., Berant, J., & Levy, O. (2021) · Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing · doi:10.18653/v1/2021.emnlp-main.446 · Cited by 353
Cited by
Walter, Y. (2024) · International Journal of Educational Technology in Higher Education · doi:10.1186/s41239-024-00448-3 · Cited by 1,113
The Hydra Effect: Emergent Self-repair in Language Model Computations
McGrath, T., Rahtz, M., Kramar, J., Mikulik, V., & Legg, S. (2023) · arXiv · doi:10.48550/arxiv.2307.15771 · Cited by 140
Interpretable machine learning for weather and climate prediction: A review
Yang, R., Hu, J., Li, Z., Mu, J., Yu, T., Xia, J., Li, X., Dasgupta, A., & Xiong, H. (2024) · Atmospheric Environment · doi:10.1016/j.atmosenv.2024.120797 · Cited by 139
Language Model Crossover: Variation through Few-Shot Prompting
Meyerson, E., Nelson, M. J., Bradley, H., Gaier, A., Moradi, A., Hoover, A. K., & Lehman, J. (2024) · ACM Transactions on Evolutionary Learning and Optimization · doi:10.1145/3694791 · Cited by 75
Agentic Large Language Models, a Survey
Plaat, A., Van Duijn, M., Van Stein, N., Preuss, M., Van der Putten, P., & Batenburg, K. J. (2025) · Journal of Artificial Intelligence Research · doi:10.1613/jair.1.18675 · Cited by 75
Metadata
- Authors
- Conmy, A., Mavor-Parker, A. N., Lynch, A., Heimersheim, S., & Garriga-Alonso, A.
- Venue
- neural information processing systems
- Published
- 2023-01-01
- Pages
- 16318-16352
- ISSN
- 1049-5258 (ISSN-L)
- OpenAlex
- W7133218086
- Cited by
- 67
- References
- 15
- FWCI
- 12.67
- Type
- conference paper
- Language
- en
- Open access
- green
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
How to cite
Cite CoChat for the checks on this page, and the paper for its own findings.
- This page
- CoChat. (2026). Citation check: Towards Automated Circuit Discovery for Mechanistic Interpretability. https://app.cochat.ai/papers/10.52202/075280-0719
- The paper
- Conmy, A., Mavor-Parker, A. N., Lynch, A., Heimersheim, S., & Garriga-Alonso, A. (2023). Towards Automated Circuit Discovery for Mechanistic Interpretability. neural information processing systems. https://doi.org/10.52202/075280-0719
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.