Intrinsic Entropy of Context Length Scaling in LLMs

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Shi, Jingzhe, Ma, Qinwei, Liu, Hongyi, Zhao, Hang, Hwang, Jeng-Neng, Li, Lei
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915826265751552
author Shi, Jingzhe
Ma, Qinwei
Liu, Hongyi
Zhao, Hang
Hwang, Jeng-Neng
Li, Lei
author_facet Shi, Jingzhe
Ma, Qinwei
Liu, Hongyi
Zhao, Hang
Hwang, Jeng-Neng
Li, Lei
contents Long Context Language Models have drawn great attention in the past few years. There has been work discussing the impact of long context on Language Model performance: some find that long irrelevant context could harm performance, while some experimentally summarize loss reduction by relevant long context as Scaling Laws. This calls for a more thorough understanding of how long context impacts Language Modeling. In this work, we (1) propose to use `Intrinsic Entropy' for explaining the impact of context length on language modeling; and (2) conduct experiments on natural language and synthetic data, validating our proposed theoretical assumptions and deductions. Our theoretical framework can provide practical insights such as establishing that training dataset size dictates an optimal context length and bounds context length scaling for certain cases. We hope our work may inspire new long context Language Models, as well as future work studying the physics of Language Models.
format Preprint
id arxiv_https___arxiv_org_abs_2502_01481
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Intrinsic Entropy of Context Length Scaling in LLMs
Shi, Jingzhe
Ma, Qinwei
Liu, Hongyi
Zhao, Hang
Hwang, Jeng-Neng
Li, Lei
Machine Learning
Computation and Language
Long Context Language Models have drawn great attention in the past few years. There has been work discussing the impact of long context on Language Model performance: some find that long irrelevant context could harm performance, while some experimentally summarize loss reduction by relevant long context as Scaling Laws. This calls for a more thorough understanding of how long context impacts Language Modeling. In this work, we (1) propose to use `Intrinsic Entropy' for explaining the impact of context length on language modeling; and (2) conduct experiments on natural language and synthetic data, validating our proposed theoretical assumptions and deductions. Our theoretical framework can provide practical insights such as establishing that training dataset size dictates an optimal context length and bounds context length scaling for certain cases. We hope our work may inspire new long context Language Models, as well as future work studying the physics of Language Models.
title Intrinsic Entropy of Context Length Scaling in LLMs
topic Machine Learning
Computation and Language
url https://arxiv.org/abs/2502.01481