AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

How Do Transformers Learn to Associate Tokens: Gradient Leading Terms Bring Mechanistic Interpretability

Shawn Im, Changdae Oh, Zhen Fang, Sharon Li

ICLR 2026 · Architectures and efficiency › Attention and transformers · approach: Attention head mechanisms · 1 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
keep output weight leading term (vs removing it) (ablation)3-layer transformer on TinyStories · TinyStories (lr 0.05)loss: 8.287 -> 5.349Improved
keep value layer 1 leading term (vs removing it) (ablation)3-layer transformer on TinyStories · TinyStories (lr 0.05)loss: 6.526 -> 5.349Improved
keep value layer 0 leading term (vs removing it) (ablation)3-layer transformer on TinyStories · TinyStories (lr 0.05)loss: 6.192 -> 5.349Improved
keep attention layer 2 leading term (vs removing it) (ablation)3-layer transformer on TinyStories · TinyStories (lr 0.05)loss: 5.361 -> 5.349No clear effect
keep attention layer 0 leading term (vs removing it) (ablation)3-layer transformer on TinyStories · TinyStories (lr 0.05)loss: 5.350 -> 5.349No clear effect
train with BPE tokenization instead of word-level (ablation)3-layer transformer on TinyStories · TinyStories (BPE 10k, lr 0.005)min cosine similarity: 0.998486 -> 0.997891No clear effect
use larger learning rate 0.05 instead of 0.005 (ablation)3-layer transformer on TinyStories · TinyStoriescosine similarity: 0.7Mixed
remove MLP output when computing embedding correlations (E_attn) (ablation)3-layer transformer on TinyStories · Pythia-1.4B (OpenWebText)cosine similarity of covariancesNo 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