paper-with-me

홈 › Papers

Pretrain-Test Task Alignment Governs Generalization in In-Context Learning

2025-09-30 · Mary I. Letey, Jacob A. Zavatone-Veth, Yue M. Lu, Cengiz Pehlevan arxiv

In-context learning (ICL) is a central capability of Transformer models, but the structures in data that enable its emergence and govern its robustness remain poorly understood. In this work, we study how the structure of pretraining tasks governs generalization in ICL. Using a solvable model for ICL of linear regression by linear attention, we derive an exact expression for ICL generalization error in high dimensions under arbitrary pretraining-testing task covariance mismatch. This leads to a new alignment measure that quantifies how much information about the pretraining task distribution is useful for inference at test time. We show that this measure directly predicts ICL performance not only in the solvable model but also in nonlinear Transformers. Our analysis further reveals a tradeoff between specialization and generalization in ICL: depending on task distribution alignment, increasing pretraining task diversity can either improve or harm test performance. Together, these results identify train-test task alignment as a key determinant of generalization in ICL.

📄 PDF Abstract BibTeX arXiv:2509.26551

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

An Effective Gram Matrix Characterizes Generalization in Deep Networks

2025-04-23 · Rubing Yang, Pratik Chaudhari

We derive a differential equation that governs the evolution of the generalization gap when a deep network is trained by gradient descent. This differential equation is controlled by two quantities, a contraction factor …

image-classificationImage Classification

Small Vocabularies, Big Gains: Pretraining and Tokenization in Time Series Models

2025-11-06 · Alexis Roger, Gwen Legate, Kashif Rasul, Yuriy Nevmyvaka 외 arxiv

Tokenization and transfer learning are two critical components in building state of the art time series foundation models for forecasting. In this work, we systematically study the effect of tokenizer design, specificall…

Representation LearningTransfer Learning

Is Large-Scale Pretraining the Secret to Good Domain Generalization?

2024-12-03 · Piotr Teterwak, Kuniaki Saito, Theodoros Tsiligkaridis, Bryan A. Plummer 외

Multi-Source Domain Generalization (DG) is the task of training on multiple source domains and achieving high classification performance on unseen target domains. Recent methods combine robust features from web-scale pre…

Domain Generalization

Domain Generalization using Pretrained Models without Fine-tuning

2022-03-09 · Ziyue Li, Kan Ren, Xinyang Jiang, Bo Li 외

Fine-tuning pretrained models is a common practice in domain generalization (DG) tasks. However, fine-tuning is usually computationally expensive due to the ever-growing size of pretrained models. More importantly, it ma…

Domain GeneralizationEnsemble Learning

Enhancing Compositional Generalization via Compositional Feature Alignment

2024-02-05 · Haoxiang Wang, Haozhe Si, Huajie Shao, Han Zhao

Real-world applications of machine learning models often confront data distribution shifts, wherein discrepancies exist between the training and test data distributions. In the common multi-domain multi-class setup, as t…