Rules Amortize, Pairings Don't: Linguistic Structure Determines What Latent Task Representations Can Replace In-Context Learning

arXiv:2610.00526v1 Announce Type: new Abstract: In-context learning (ICL) can be amortized into latent objects (task vectors, function vectors, context vectors) that recover few-shot behavior at zero-shot inference cost, but recent theory shows a static vector acts as a single synthetic demonstrati…

Read the full story at arXiv cs.CL ↗

Timeline · 1 report

1. 2026-10-02 04:00 · arXiv cs.CL
   Rules Amortize, Pairings Don't: Linguistic Structure Determines What Latent Task Representations Can Replace In-Context Learning