Saved in:
Bibliographic Details
Main Authors: Buonanno, Amedeo, Rivetti, Alessandro, Palmieri, Francesco A. N., Di Gennaro, Giovanni, Romano, Gianmarco
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2507.15347
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916870514278400
author Buonanno, Amedeo
Rivetti, Alessandro
Palmieri, Francesco A. N.
Di Gennaro, Giovanni
Romano, Gianmarco
author_facet Buonanno, Amedeo
Rivetti, Alessandro
Palmieri, Francesco A. N.
Di Gennaro, Giovanni
Romano, Gianmarco
contents This work explores entropy analysis as a tool for probing information distribution within Transformer-based architectures. By quantifying token-level uncertainty and examining entropy patterns across different stages of processing, we aim to investigate how information is managed and transformed within these models. As a case study, we apply the methodology to a GPT-based large language model, illustrating its potential to reveal insights into model behavior and internal representations. This approach may offer insights into model behavior and contribute to the development of interpretability and evaluation frameworks for transformer-based models
format Preprint
id arxiv_https___arxiv_org_abs_2507_15347
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Probing Information Distribution in Transformer Architectures through Entropy Analysis
Buonanno, Amedeo
Rivetti, Alessandro
Palmieri, Francesco A. N.
Di Gennaro, Giovanni
Romano, Gianmarco
Computation and Language
Machine Learning
This work explores entropy analysis as a tool for probing information distribution within Transformer-based architectures. By quantifying token-level uncertainty and examining entropy patterns across different stages of processing, we aim to investigate how information is managed and transformed within these models. As a case study, we apply the methodology to a GPT-based large language model, illustrating its potential to reveal insights into model behavior and internal representations. This approach may offer insights into model behavior and contribute to the development of interpretability and evaluation frameworks for transformer-based models
title Probing Information Distribution in Transformer Architectures through Entropy Analysis
topic Computation and Language
Machine Learning
url https://arxiv.org/abs/2507.15347