AI Paper Atlas

Research map / Architectures and efficiency / RNNs and normalization

Feature Resemblance: Towards a Theoretical Understanding of Analogical Reasoning in Transformers

Ruichen Xu, Wenjing Yan, Ying-Jun Angela Zhang

ICML 2026 · Architectures and efficiency › RNNs and normalization · approach: Layer-wise analysis of language models · 0 citations (OpenAlex)

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
learn similarity premises before attribution premises (curriculum) instead of reverse (ablation)GPT-2 trained from scratch on synthetic triples · Synthetic analogical reasoningsuccess rate: 0.7% -> 100%Improved
learn similarity premises before attribution premises (ablation)One-layer transformer · Synthetic analogical reasoningsuccess rate: 0% -> 100%Improved
joint training on similarity and attribution premisesGPT-2 · Synthetic analogical reasoningfeature similarity: 0.995±6e-4 with 100% successImproved
add identity bridge examples (b -> b) to two-hop training data (ablation)GPT-2 · Synthetic two-hop reasoningsuccess rate: 0% -> 99.7%Improved
add identity bridge examples (ablation)One-layer transformer · Synthetic two-hop reasoningsuccess rate: 1.7% -> 100%Improved
add identity bridge examples (ablation)GPT-2 · Synthetic two-hop reasoningfeature similarity: 0.016 -> 0.9642Improved
learn similarity before attribution (late on attribution vs late on similarity) (ablation)Llama-3-1B · Natural-language analogysuccess rate: 53% -> 86%Improved
learn similarity before attribution (ablation)Qwen-2.5-1.5B · Natural-language analogysuccess rate: 33% -> 72%Improved
learn similarity before attribution (ablation)Qwen-3-8B · Natural-language analogysuccess rate: 34% -> 75%Improved
joint training instead of late-on-attribution (ablation)Llama-3-1B · Natural-language analogysuccess rate: 86% -> 73%Hurt
compare joint vs late-on-attribution (ablation)Qwen-2.5-1.5B and Qwen-3-8B · Natural-language analogysuccess rate: 85% vs 72% (Qwen-2.5-1.5B); 84% vs 75% (Qwen-3-8B)No clear effect

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