SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2602.00747 · ICML · 2026

Decouple Searching from Training: Scaling Data Mixing via Model Merging for Large Language Model Pre-training

Yao Hu, Shaosheng Cao, Fei Zhao, Jieying Ye, Zheyong Xie, Kaiyan Zhao, Haifeng Liu, Shengrui Li, Fangcheng Shi

arXiv · PDF · Open in the Atlas

Code that ran

7 of 7 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; 0 honoured the contract we drafted, 1 violated it and 0 have no contract verdict; 6 failures of our instrument; 0 no recorded run), measured 2026-11-22; 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/2602.00747.json · badge: /api/ran/2602.00747.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
Lucius-lsr/DeMix canonical (classifier label)
classifier basis: census
7 ran · 0 no recorded run
FunctionStatusWhere it lives
eval Ran
ran; our contract draft was wrong, not the code
Lucius-lsr/DeMix/eval_merged/proxy_eval.py
pointer only (licence: NONE) · get_code("e3aff3e98dea865c")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, Lucius-lsr/DeMix 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.

filter_by_cosine_similarity Ran
ran, violated the contract we drafted
Lucius-lsr/DeMix/iterative_sample/sample.py
pointer only (licence: NONE) · get_code("51bd50a70406e33c")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, Lucius-lsr/DeMix 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.

generate_weights_dirichlet Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
Lucius-lsr/DeMix/iterative_sample/sample.py
pointer only (licence: NONE) · get_code("51231590fece73f8")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, Lucius-lsr/DeMix 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_benchmark_data Ran
ran; our contract draft was wrong, not the code
Lucius-lsr/DeMix/eval_merged/proxy_eval.py
pointer only (licence: NONE) · get_code("5800df7ee414bdc8")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, Lucius-lsr/DeMix 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_top_spearman Ran
ran; our contract draft was wrong, not the code
Lucius-lsr/DeMix/eval_merged/proxy_eval.py
pointer only (licence: NONE) · get_code("ab050c4e7ddda67e")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, Lucius-lsr/DeMix 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_y Ran
ran; our contract draft was wrong, not the code
Lucius-lsr/DeMix/iterative_sample/train_predictor.py
pointer only (licence: NONE) · get_code("497e8132df0c6503")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, Lucius-lsr/DeMix 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.

sort_and_deduplicate Ran
ran; the contract check could not run (RAISES, DEP_MISSING or TIMEOUT)
Lucius-lsr/DeMix/iterative_sample/sample.py
pointer only (licence: NONE) · get_code("ad93b5075adbf5a9")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, Lucius-lsr/DeMix 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.

Also present in other papers' code

Of 7 samples lifted from this paper's linked repositories, 7 carry a function identity and 0 do not. Of the 7 distinct functions with an identity, 1 are also present in other papers whose code we hold (1 with at least one paper dated earlier, 0 the same month, 0 later) and 6 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
sort_and_deduplicate 1 other paper
1 dated earlier
2407.01492 (2024-07)
RegMix: Data Mixture as Regression for Language Model Pre-training

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

Determining an effective data mixture is a key factor in Large Language Model (LLM) pre-training, where models must balance general competence with proficiency on hard tasks such as math and code. However, identifying an optimal mixture remains an open challenge, as existing approaches either rely on unreliable tiny-scale proxy experiments or require prohibitively expensive largescale exploration. To address this, we propose Decouple Searching from Training Mix (DeMix), a novel framework that leverages model merging to predict optimal data ratios. Instead of training proxy models for every sampled mixture, DeMix trains component models on candidate datasets at scale and derives data mixture proxies via weighted model merging. This paradigm decouples search from training costs, enabling evaluation of unlimited sampled mixtures without extra training burden and thus facilitating better mixture discovery through more search trials. Extensive experiments demonstrate that DeMix breaks the trade-off between sufficiency, accuracy and efficiency, obtaining the optimal mixture with higher benchmark performance at lower search cost. Additionally, we release the DeMix Corpora, a comprehensive 22T-token dataset comprising high-quality pre-training data with validated mixtures to facilitate open research. Our code and DeMix Corpora is available at https: //github.com/Lucius-lsr/DeMix.

For agents

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

get_harvested_code_for_paper("2602.00747")
get_code_for_paper("2602.00747")
have("2602.00747")

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

curl https://syntology.ai/api/ran/2602.00747.json

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

[![Syntology run record](https://syntology.ai/api/ran/2602.00747.svg)](https://syntology.ai/paper/2602.00747)

Connect an agent — have() is free.