Transferring Linear Features Across Language Models With Model Stitching

NeurIPSSpotlight2025

Authors
Alan Chen, Jack Merullo, Alessandro Stolfo, Ellie Pavlick
Venue
NeurIPS 2025
Track
Spotlight

TL;DR

In this work, we demonstrate that affine mappings between residual streams of language models is a cheap way to effectively transfer represented features between models…

Opening excerpt from the authors’ abstract. source

Read the paper

Topics

language model

← All NeurIPS 2025 Spotlight papers · Browse the whole archive