AI Paper Atlas

Research map / Architectures and efficiency / Attention and transformers

Data Distributional Properties Drive Emergent In-Context Learning in Transformers

Stephanie C.Y. Chan, Adam Santoro, Andrew Kyle Lampinen, Jane X Wang, Aaditya K Singh, Pierre Harvey Richemond, James McClelland, Felix Hill

NeurIPS 2022 · Architectures and efficiency › Attention and transformers · approach: Transformer block and depth design · 50 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
Increase the proportion of bursty training sequences (ablation)12-layer transformer · Omniglot holdout classesin-context learning accuracyImproved
Increase the proportion of bursty training sequences (ablation)12-layer transformer · Omniglot trained classesin-weights learningHurt
Increase training classes from 100 to 1600 (ablation)Transformer trained with burstiness fixed at p(bursty)=0.9 · Omniglot holdout classesin-context learning accuracyImproved
Increase training classes from 100 to 1600 (ablation)Transformer trained with burstiness fixed at p(bursty)=0.9 · Omniglot trained classesin-weights learningHurt
Expand training classes to 12800 using rotated and flipped images (ablation)Transformer trained with p(bursty)=0.9 · Omniglot holdout classesin-context learning accuracyImproved
Expand training classes to 12800 using rotated and flipped images (ablation)Transformer trained with p(bursty)=0.9 · Omniglot trained classesin-weights learningHurt
Increase the number of possible labels per image class (ablation)Transformer trained with p(bursty)=0.9 · Omniglot holdout classesin-context learning accuracyImproved
Increase within-class image variation (ablation)Transformer · Omniglot holdout classesin-context learning accuracyImproved
Increase within-class image variation (ablation)Transformer · Omniglot trained classesin-weights learningHurt
Train on a Zipfian class distribution with exponent 1 (ablation)Transformer trained with 12800 classes and p(bursty)=0.9 · Omniglot holdout classesin-context learning accuracyImproved
Train on a Zipfian class distribution with exponent 1 (ablation)Transformer trained with 12800 classes and p(bursty)=0.9 · Omniglot common classesin-weights learningImproved
Increase Zipfian skew beyond the uniform distribution (ablation)Transformer trained with 12800 classes and p(bursty)=0.9 · Omniglot holdout classesin-context learning 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