The Garden of Forking Paths: Observing Dynamic Parameters Distribution in Large Language Models

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Nicolini, Carlo, Staiano, Jacopo, Lepri, Bruno, Marino, Raffaele
Format: Preprint
Published: 2024
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909135720677376
author Nicolini, Carlo
Staiano, Jacopo
Lepri, Bruno
Marino, Raffaele
author_facet Nicolini, Carlo
Staiano, Jacopo
Lepri, Bruno
Marino, Raffaele
contents A substantial gap persists in understanding the reasons behind the exceptional performance of the Transformer architecture in NLP. A particularly unexplored area involves the mechanistic description of how the distribution of parameters evolves over time during training. In this work we suggest that looking at the time evolution of the statistic distribution of model parameters, and specifically at bifurcation effects, can help understanding the model quality, potentially reducing training costs and evaluation efforts and empirically showing the reasons behind the effectiveness of weights sparsification.
format Preprint
id arxiv_https___arxiv_org_abs_2403_08739
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle The Garden of Forking Paths: Observing Dynamic Parameters Distribution in Large Language Models
Nicolini, Carlo
Staiano, Jacopo
Lepri, Bruno
Marino, Raffaele
Computation and Language
Disordered Systems and Neural Networks
Statistical Mechanics
Artificial Intelligence
A substantial gap persists in understanding the reasons behind the exceptional performance of the Transformer architecture in NLP. A particularly unexplored area involves the mechanistic description of how the distribution of parameters evolves over time during training. In this work we suggest that looking at the time evolution of the statistic distribution of model parameters, and specifically at bifurcation effects, can help understanding the model quality, potentially reducing training costs and evaluation efforts and empirically showing the reasons behind the effectiveness of weights sparsification.
title The Garden of Forking Paths: Observing Dynamic Parameters Distribution in Large Language Models
topic Computation and Language
Disordered Systems and Neural Networks
Statistical Mechanics
Artificial Intelligence
url https://arxiv.org/abs/2403.08739