AI Paper Atlas

Research map / Architectures and efficiency / RNNs and normalization

The Geometry and Topology of Circuits: the Manifolds of Modular Addition

Gabriela Moisescu-Pareja, Gavin McCracken, Harley Wiltzer, Colin Daniels, Vincent Létourneau, Doina Precup, Jonathan Love

ICLR 2026 · Architectures and efficiency › RNNs and normalization · approach: Mechanistic circuit interpretability

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
switch from uniform attention (Pizza) to trainable attention (Clock)modular addition transformer · layer-1 PCA geometryvariance explained by top 2 PCsNo clear effect
switch from uniform attention to trainable attentionmodular addition transformer · phase alignment distribution (max activation)MMD: MMD 0.0338No clear effect
concatenate embeddings (MLP-Concat) instead of adding (MLP-Add)MLP on modular addition · layer-1 manifoldCKA to torus vs disc: disc CKA 0.707 vs torus CKA 0.994Improved
change architecture (MLP-Add, Attention 0.0/1.0, MLP-Concat)modular addition networks · logit manifoldCKA to circle: 0.926 to 0.986No clear effect
increase depth from 1 to 2-3 layersMLP-Add and Attention · Betti numbers per layertopologyMixed
use Zhong et al. diagonal-dependence as Pizza signatureClock vs Pizza distinction · activation dependence on a-bactivation patternHurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy