SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2407.06645 · 2024

Entropy Law: The Story Behind Data Compression and LLM Performance

arXiv · PDF · Open in the Atlas

Code that ran

4 of 5 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (0 of them constructed an object rather than computing a result; 2 honoured the contract we drafted, 1 violated it and 0 have no contract verdict; 1 failure of our instrument; 1 no recorded run), measured 2026-11-22; repository links from the Papers with Code archive attributed under CC BY-SA, no affiliation; bodies inline only under a permissive licence as GitHub detects it. Not a reproduction of the paper's results.

"Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results. This record, dated, as JSON: /api/ran/2407.06645.json · badge: /api/ran/2407.06645.svg

Each repository's role is a label: a classifier's output recorded on the link, or, where no role is recorded, what the Papers with Code archive marked. The line under it names its basis. Neither says who wrote the code.

RepositoryRoleRan
ustc-starteam/zip canonical (classifier label)
classifier basis not recorded on the link
2 ran · 1 no recorded run
MindCode-4/code-11 community (classifier label)
classifier basis not recorded on the link
2 ran · 0 no recorded run
FunctionStatusWhere it lives
exists Ran
ran, violated the contract we drafted
MindCode-4/code-11/Adaptive-Nesterov-Momentum-Algorithm/adan.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from ThomasMrY/VCT, is pointer only (licence: NONE) · get_code("aa5486a3650902d8")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from ThomasMrY/VCT, and GitHub detected no licence in that repository when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

get_compression_ratio Ran
ran, honoured the contract we drafted
ustc-starteam/zip/ZIP.py
pointer only (licence: MIT); the stored copy get_code serves, from xiatingyu/sft-dataselection-at-scale, is pointer only (licence: NONE) · get_code("b2b37eed6f3f95fb")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from xiatingyu/sft-dataselection-at-scale, and GitHub detected no licence in that repository when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

load_original_data_pool Ran
ran; our contract draft was wrong, not the code
ustc-starteam/zip/ZIP.py
pointer only (licence: MIT); the stored copy get_code serves, from xiatingyu/sft-dataselection-at-scale, is pointer only (licence: NONE) · get_code("be0b30b9715eb548")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from xiatingyu/sft-dataselection-at-scale, and GitHub detected no licence in that repository when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

noop Ran
ran, honoured the contract we drafted
MindCode-4/code-11/Fixup-Initialization/model_build.py
pointer only (licence: Apache-2.0); the stored copy get_code serves, from Abhimanyu08/Fixup_Initialization, is pointer only (licence: NONE) · get_code("5e2ed23acb71ee11")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (Apache-2.0), but get_code serves the stored copy, first harvested from Abhimanyu08/Fixup_Initialization, and GitHub detected no licence in that repository when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. A run is recorded.

selec_data_from_corpus Not yet run
unverified (harvested, no recorded run)
ustc-starteam/zip/ZIP.py
pointer only (licence: MIT); the stored copy get_code serves, from xiatingyu/sft-dataselection-at-scale, is pointer only (licence: NONE) · get_code("86d72ce381c39d49")
How it ran

The body of this sample is withheld: this paper's link allows the body inline (MIT), but get_code serves the stored copy, first harvested from xiatingyu/sft-dataselection-at-scale, and GitHub detected no licence in that repository when it was checked. The setup block that called it is withheld with it, because it is written against that body. source_path and origin say where the code lives, to read under the repository's own terms. No invocation recorded.

Also present in other papers' code

Of 5 samples lifted from this paper's linked repositories, 5 carry a function identity and 0 do not. Of the 5 distinct functions with an identity, 5 are also present in other papers whose code we hold (2 with at least one paper dated earlier, 2 the same month, 5 later) and 0 are only here among the 52,977 papers whose code Syntology holds (as of 2026-09-26).

Only here means: among the 52,977 papers whose code Syntology holds (51,887 with lifted samples, 1,090 through their assemblies only), as of 2026-09-26, no other paper's lifted samples or current assembly bricks contain an identical function body after normalisation.

FunctionAlso present inEarliest-datedOthers named
exists 486 other papers
362 dated earlier, 8 the same month, 105 later, 9 same year, order not determinable, 2 undated · 78 only as an assembly brick
1603.08155 (2016-03)
Perceptual Losses for Real-Time Style Transfer and Super-Resolution
1611.07004 (2016-11)
Image-to-Image Translation with Conditional Adversarial Networks
1804.04637 (2018-04)
EMBER: An Open Dataset for Training Static PE Malware Machine Learning Models
and 483 more not named here
noop 9 other papers
7 dated earlier, 1 the same month, 1 later
1805.08318 (2018-05)
Self-Attention Generative Adversarial Networks
1901.09321 (2019-01)
Fixup Initialization: Residual Learning Without Normalization
1906.03028 (2019-06)
Automatic Reparameterisation of Probabilistic Programs
and 6 more not named here
get_compression_ratio 1 other paper
1 later
2410.09335 (2024-10)
Rethinking Data Selection at Scale: Random Selection is Almost All You Need
load_original_data_pool 1 other paper
1 later
2410.09335 (2024-10)
Rethinking Data Selection at Scale: Random Selection is Almost All You Need
selec_data_from_corpus 1 other paper
1 later
2410.09335 (2024-10)
Rethinking Data Selection at Scale: Random Selection is Almost All You Need

Presence, not direction: an identical function body in two papers' repositories does not say who wrote it, which came from which, or that either paper's method rests on it. Matching is exact after normalisation, so a renamed or edited function is not matched. Dates are when papers appeared, not when code was written. Only .py files the harvester lifted are covered; notebooks are not read. A new-style arXiv id YYMM.NNNN dates a paper to the month 20YY-MM; otherwise Paper.year dates it to the year; otherwise it is undated. Dates are when the paper appeared, not when the code was written. Read 2026-11-22.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Data is the cornerstone of large language models (LLMs), but not all data is useful for model learning. Carefully selected data can better elicit the capabilities of LLMs with much less computational overhead. Most methods concentrate on evaluating the quality of individual samples in data selection, while the combinatorial effects among samples are neglected. Even if each sample is of perfect quality, their combinations may be suboptimal in teaching LLMs due to their intrinsic homogeneity or contradiction. In this paper, we aim to uncover the underlying relationships between LLM performance and data selection. Inspired by the information compression nature of LLMs, we uncover an ``entropy law'' that connects LLM performance with data compression ratio and first-epoch training loss, which reflect the information redundancy of a dataset and the mastery of inherent knowledge encoded in this dataset, respectively. Through both theoretical deduction and empirical evaluation, we find that model performance is negatively correlated to the compression ratio of training data, which usually yields a lower training loss. Based on the findings of the entropy law, we propose a quite efficient and universal data selection method named \textbf{ZIP} for training LLMs, which aim to prioritize data subsets exhibiting a low compression ratio. Based on a multi-stage algorithm that selects diverse data in a greedy manner, we can obtain a good data subset with satisfactory diversity. Extensive experiments have been conducted to validate the entropy law and the superiority of ZIP across different LLM backbones and alignment stages. We also present an interesting application of entropy law that can detect potential performance risks at the beginning of model training.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2407.06645")
get_code_for_paper("2407.06645")
have("2407.06645")

The run record, dated, one paper per request, free:

curl https://syntology.ai/api/ran/2407.06645.json

A badge for a README (the split and the date, never a ratio):

[![Syntology run record](https://syntology.ai/api/ran/2407.06645.svg)](https://syntology.ai/paper/2407.06645)

Connect an agent — have() is free.