View a PDF of the paper titled Emergence of a High-Dimensional Abstraction Phase in Language Transformers, by Emily Cheng and Diego Doimo and Corentin Kervadec and Iuri Macocco and Jade Yu and Alessandro Laio and Marco Baroni
Abstract:A language model (LM) is a mapping from a linguistic context to an output token. However, much remains to be known about this mapping, including how its geometric properties relate to its function. We take a high-level geometric approach to its analysis, observing, across five pre-trained transformer-based LMs and three input datasets, a distinct phase characterized by high intrinsic dimensionality. During this phase, representations (1) correspond to the first full linguistic abstraction of the input; (2) are the first to viably transfer to downstream tasks; (3) predict each other across different LMs. Moreover, we find that an earlier onset of the phase strongly predicts better language modelling performance. In short, our results suggest that a central high-dimensionality phase underlies core linguistic processing in many common LM architectures.
Submission history
From: Emily Cheng [view email]
[v1]
Fri, 24 May 2024 11:49:07 UTC (8,818 KB)
[v2]
Fri, 20 Dec 2024 23:09:46 UTC (9,946 KB)
Source link
lol