AI Paper Atlas

Research map / Large language models / Language, safety and interpretability

Emergent Analogical Reasoning in Transformers

Gouki Minegishi, Jingyuan Feng, Hiroki Furuta, Takeshi Kojima, Yusuke Iwasawa, Yutaka Matsuo

ICML 2026 · Large language models › Language, safety and interpretability · approach: Reasoning mechanisms and failure modes · 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
reduce number of relations to 100 (ablation)1-layer Transformer · Synthetic analogy task, 1-layer TransformerOOD analogical accuracy: analogical reasoning fails at |R| = 100Hurt
increase number of entities (ablation)1-layer Transformer · Synthetic analogy task, 1-layer Transformertime to acquire analogical reasoningHurt
raise analogical OOD ratio to 0.9 (ablation)1-layer Transformer · Synthetic analogy task, 1-layer TransformerOOD analogical accuracyHurt
raise compositional OOD ratio (ablation)1-layer Transformer · Synthetic analogy task, 1-layer Transformeranalogical reasoningNo clear effect
raise weight decay from 0 to 0.01 or 0.1 (ablation)1-layer Transformer · Synthetic analogy task, 1-layer Transformeremergence time of analogical reasoningImproved
raise weight decay to 1.0 (ablation)1-layer Transformer · Synthetic analogy task, 1-layer TransformerOOD analogical accuracy: fails to be learned; compositional reasoning remains robustHurt
increase batch size (ablation)1-layer Transformer · Synthetic analogy task, 1-layer Transformeremergence time of analogical reasoningImproved
increase width from 64 to 128 or 256 (ablation)Transformer · Synthetic analogy task, 1-layer TransformerOOD analogical accuracyImproved
increase width to 512 (ablation)Transformer · Synthetic analogy task, 1-layer TransformerOOD analogical accuracyHurt
increase model size (ablation)Transformer · Synthetic analogy task, 1-layer Transformercompositional reasoning accuracyImproved
increase depth to 2 or 4 layers (fixed default optimization) (ablation)Transformer · Synthetic analogy task, 1-layer TransformerOOD analogical accuracyHurt
add noise to functor-transferred relations (imperfect isomorphism) (ablation)1-layer Transformer · Synthetic analogy task, 1-layer TransformerOOD analogical accuracyHurt

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