AI Paper Atlas

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

Predicting the Emergence of Induction Heads in Language Model Pretraining

Tatsuya Aoyama, Ethan Gotlieb Wilcox, Nathan Schneider

ICML 2026 · Large language models › Language, safety and interpretability · approach: Reasoning mechanisms and failure modes

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
increase batch size (fixed context 1024) (ablation)GPT2 50M · Natural data (CC100)induction head emergence point (updates): rho=-1.0, p<0.001Improved
increase batch size (ablation)GPT2 50M · Natural data (CC100)final induction head strength (PS): larger batch -> lower eventual PSHurt
increase context size (fixed batch 16) (ablation)GPT2 50M · Natural data (CC100)induction head emergence point: rho=-1.0, p<0.001Improved
increase context size (ablation)GPT2 50M · Natural data (CC100)slope after emergence: rho=1.0, p<0.001Improved
increase proportion of chunks with bigram repetitions (ablation)GPT2 50M · Natural data (CC100)emergence point: rho=-0.64, p=0.12No clear effect
increase proportion of chunks with bigram repetitions (ablation)GPT2 50M · Natural data (CC100)slope: rho=0.93, p=0.002Improved
add model size as predictor of emergence updates (ablation)log-linear law · Natural data (CC100)emergence updates regression: -0.0027, p 0.966No clear effect
make next-token distribution independent of current token (-LD) (ablation)synthetic Markov data · Synthetic datainduction head formationHurt
use Zipfian marginal with categoriality at near-boundary repetition (0.1-0.3) (ablation)synthetic Markov data · Synthetic datainduction head formationImproved
increase batch size beyond B=64 (context 1024) (ablation)GPT2 50M · Natural data (CC100)final validation loss: lower until B=64, higher beyondMixed

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