Understanding Prompt Tuning and In-Context Learning via Meta-Learning

NeurIPSSpotlight2025

Authors: Tim Genewein, Li Kevin Wenliang, Jordi Grau-Moya, Anian Ruoss, Laurent Orseau, Marcus Hutter
Venue: NeurIPS 2025
Track: Spotlight

TL;DR

Prompting is one of the main ways to adapt a pretrained model to target tasks…

Opening excerpt from the authors’ abstract. source

Read the paper

OpenReview

Topics

in-context learning meta-learning

← All NeurIPS 2025 Spotlight papers · Browse the whole archive