SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2602.08351 · 2026

The Chicken and Egg Dilemma: Co-optimizing Data and Model Configurations for LLMs

Bryan Kian, Hsiang Low, Nancy Chen, Apivich Hemachandra, Chuan-Sheng Foo, Gregory Kang, Ruey Lau, Zhiliang Chen, Zhengyuan Liu, Alfred Wei, Lun Leong, Shao Ong, and 1 more

arXiv · PDF · Open in the Atlas

Code that ran

6 of 11 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; 3 honoured the contract we drafted, 0 violated it and 0 have no contract verdict; 3 failures of our instrument; 5 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/2602.08351.json · badge: /api/ran/2602.08351.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
princeton-nlp/LESS canonical (classifier label)
classifier basis: census
5 ran · 3 no recorded run
ruz048/AutoLoRA canonical (classifier label)
classifier basis: census
1 ran · 2 no recorded run
FunctionStatusWhere it lives
calculate_influence_score Ran
ran; our contract draft was wrong, not the code
princeton-nlp/LESS/less/data_selection/matching.py
code served (permissive licence) · get_code("8553c4772e3764ad")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned Tensor, ndim 2, dtype family float, shape (2, 3) as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import torch

torch.manual_seed(0)

# Create training_info tensor: shape (N, N_DIM) = (2, 4)
training_info = torch.randn(2, 4)

# Create validation_info tensor: shape (N_VALID, N_DIM) = (3, 4)
validation_info = torch.randn(3, 4)

ARGS = {
    "training_info": training_info,
    "validation_info": validation_info,
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
torch — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type Tensor, ndim 2, dtype family float, shape (2, 3) — shape (2, 3) as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
count_lines Ran
ran, honoured the contract we drafted
princeton-nlp/LESS/less/data_selection/write_selected_data.py
code served (permissive licence) · get_code("44e3cec9ab30944c")
How it ran

Ran in Syntology's sandbox on this synthesized input (drafted by a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API) and returned int, ndim 0, dtype family int, shape () as a later run recorded it. The image it ran in was not recorded. If your copy returns the same type, ndim and shape on these inputs, it matched what we observed; nothing here checks that. It does NOT mean the output is correct: nothing has compared it to the paper.

import tempfile
import os

# Create a temporary file with some lines of text
temp_file = tempfile.NamedTemporaryFile(mode='w', delete=False, encoding='utf-8')
temp_file.write("line 1\n")
temp_file.write("line 2\n")
temp_file.write("line 3\n")
temp_file.write("line 4\n")
temp_file.write("line 5\n")
temp_file.close()

ARGS = {
    'filename': temp_file.name
}
Setup block drafted by
a language model, Claude Haiku 4.5, prompt python_fixture_v1, through the batch API
Imports
none outside the standard library — import names as written, not package names to install; a name the repository itself provides is not a package
Returned
type int, ndim 0, dtype family int, shape () — shape () as a later run recorded it
dtype family
int means an integer dtype; float means any other dtype the instrument saw, bool and complex included (fixture_check.py records only these two families)
Environment
image not recorded; re-verify against syntology-harvest-sandbox:v5-deps; not recorded for this sample: the harness installs a cuda-to-cpu shim (cuda_shim.py) before every sample and no run records whether a sample needed it; seed: none by the harness; any seed is in the setup block
get_max_saved_index Ran
ran, honoured the contract we drafted
princeton-nlp/LESS/less/data_selection/collect_grad_reps.py
pointer only (licence: MIT); the stored copy get_code serves, from FanZT6/OptimSyn, is pointer only (licence: NONE) · get_code("ed43dd7816626e87")
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 FanZT6/OptimSyn, 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_output Ran
ran; our contract draft was wrong, not the code
princeton-nlp/LESS/less/data_selection/collect_grad_reps.py
pointer only (licence: MIT); the stored copy get_code serves, from FanZT6/OptimSyn, is pointer only (licence: NONE) · get_code("808c09bce5840bb9")
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 FanZT6/OptimSyn, 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.

lora_state_dict Ran
ran; our contract draft was wrong, not the code
ruz048/AutoLoRA/loralib/utils.py
pointer only (licence: NONE); the stored copy get_code serves, from shihaohou/FedPuReL, is pointer only (licence: NONE) · get_code("660073ae6862ac6e")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, ruz048/AutoLoRA 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.

tokenize Ran
ran, honoured the contract we drafted
princeton-nlp/LESS/less/data_selection/get_validation_dataset.py
pointer only (licence: MIT); the stored copy get_code serves, from FanZT6/OptimSyn, is pointer only (licence: NONE) · get_code("5d4f58c63866080e")
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 FanZT6/OptimSyn, 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.

encode_data Not yet run
unverified (harvested, no recorded run)
princeton-nlp/LESS/less/data_selection/get_training_dataset.py
code served (permissive licence) · get_code("25035506ff07d0a3")
How it ran

no invocation recorded.

get_data_statistics Not yet run
unverified (harvested, no recorded run)
princeton-nlp/LESS/less/train/data_arguments.py
pointer only (licence: MIT); the stored copy get_code serves, from FanZT6/OptimSyn, is pointer only (licence: NONE) · get_code("fc5b455108e53a6e")
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 FanZT6/OptimSyn, 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.

none_or_str Not yet run
unverified (harvested, no recorded run)
princeton-nlp/LESS/less/train/data_arguments.py
pointer only (licence: MIT); the stored copy get_code serves, from FanZT6/OptimSyn, is pointer only (licence: NONE) · get_code("12b6473bde25b1dd")
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 FanZT6/OptimSyn, 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.

postprocess_qa_predictions Not yet run
unverified (harvested, no recorded run)
ruz048/AutoLoRA/finetune/utils_qa.py
pointer only (licence: NONE); the stored copy get_code serves, from misonsky/chunk, is pointer only (licence: NONE) · get_code("ac56d3e7b79dbc18")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, ruz048/AutoLoRA 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.

postprocess_qa_predictions_with_beam_search Not yet run
unverified (harvested, no recorded run)
ruz048/AutoLoRA/finetune/utils_qa.py
pointer only (licence: NONE); the stored copy get_code serves, from misonsky/chunk, is pointer only (licence: NONE) · get_code("8c51f5aaf2f659e2")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, ruz048/AutoLoRA 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 11 samples lifted from this paper's linked repositories, 11 carry a function identity and 0 do not. Of the 11 distinct functions with an identity, 11 are also present in other papers whose code we hold (11 with at least one paper dated earlier, 0 the same month, 8 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
lora_state_dict 13 other papers
12 dated earlier, 1 later
2202.00535 (2022-02)
Novelty Controlled Paraphrase Generation with Retrieval Augmented Conditional Prompt Tuning
2202.07962 (2022-02)
Revisiting Parameter-Efficient Tuning: Are We Really There Yet?
2211.09790 (2022-11)
ConStruct-VL: Data-Free Continual Structured VL Concepts Learning
and 10 more not named here
postprocess_qa_predictions_with_beam_search 5 other papers
4 dated earlier, 1 later
2106.09685 (2021-06)
LoRA: Low-Rank Adaptation of Large Language Models
2205.01068 (2022-05)
OPT: Open Pre-trained Transformer Language Models
2023.acl-short.31 (2023 (year only))
PLUE: Language Understanding Evaluation Benchmark for Privacy Policies in English
and 2 more not named here
get_max_saved_index 4 other papers
3 dated earlier, 1 later
2402.04333 (2024-02)
LESS: Selecting Influential Data for Targeted Instruction Tuning
2407.15235 (2024-07)
TAGCOS: Task-agnostic Gradient Clustered Coreset Selection for Instruction Tuning Data
2411.12882 (2024-11)
ProSec: Fortifying Code LLMs with Proactive Security Alignment
and 1 more not named here
get_output 4 other papers
3 dated earlier, 1 later
2402.04333 (2024-02)
LESS: Selecting Influential Data for Targeted Instruction Tuning
2407.15235 (2024-07)
TAGCOS: Task-agnostic Gradient Clustered Coreset Selection for Instruction Tuning Data
2411.12882 (2024-11)
ProSec: Fortifying Code LLMs with Proactive Security Alignment
and 1 more not named here
postprocess_qa_predictions 4 other papers
3 dated earlier, 1 later
2106.09685 (2021-06)
LoRA: Low-Rank Adaptation of Large Language Models
2023.acl-short.31 (2023 (year only))
PLUE: Language Understanding Evaluation Benchmark for Privacy Policies in English
2310.14709 (2023-10)
Once Upon a $\textit{Time}$ in $\textit{Graph}$: Relative-Time Pretraining for Complex Temporal Reasoning
and 1 more not named here
get_data_statistics 3 other papers
2 dated earlier, 1 later
2402.04333 (2024-02)
LESS: Selecting Influential Data for Targeted Instruction Tuning
2407.15235 (2024-07)
TAGCOS: Task-agnostic Gradient Clustered Coreset Selection for Instruction Tuning Data
2604.00536 (2026-04)
Optimsyn: Influence-Guided Rubrics Optimization for Synthetic Data Generation
none_or_str 3 other papers
2 dated earlier, 1 later
2402.04333 (2024-02)
LESS: Selecting Influential Data for Targeted Instruction Tuning
2407.15235 (2024-07)
TAGCOS: Task-agnostic Gradient Clustered Coreset Selection for Instruction Tuning Data
2604.00536 (2026-04)
Optimsyn: Influence-Guided Rubrics Optimization for Synthetic Data Generation
tokenize 3 other papers
2 dated earlier, 1 later
2402.04333 (2024-02)
LESS: Selecting Influential Data for Targeted Instruction Tuning
2407.15235 (2024-07)
TAGCOS: Task-agnostic Gradient Clustered Coreset Selection for Instruction Tuning Data
2604.00536 (2026-04)
Optimsyn: Influence-Guided Rubrics Optimization for Synthetic Data Generation
encode_data 2 other papers
2 dated earlier
2402.04333 (2024-02)
LESS: Selecting Influential Data for Targeted Instruction Tuning
2407.15235 (2024-07)
TAGCOS: Task-agnostic Gradient Clustered Coreset Selection for Instruction Tuning Data
calculate_influence_score 1 other paper
1 dated earlier
2402.04333 (2024-02)
LESS: Selecting Influential Data for Targeted Instruction Tuning
count_lines 1 other paper
1 dated earlier
2402.04333 (2024-02)
LESS: Selecting Influential Data for Targeted Instruction Tuning

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

Co-optimizing data and model configurations for training LLMs presents a classic chicken-and-egg dilemma: The best training data configuration (e.g., data mixture) for a downstream task depends on the chosen model configuration (e.g., model architecture), and vice versa. However, jointly optimizing both data and model configurations is often deemed intractable, and existing methods focus on either data or model optimization without considering their interaction. We introduce JoBS, an approach that uses a scaling-law-inspired performance predictor to aid Bayesian optimization (BO) in jointly optimizing LLM training data and model configurations efficiently. JoBS allocates a portion of the optimization budget to learn an LLM performance predictor that predicts how promising a training configuration is from a small number of training steps. The remaining budget is used to perform BO entirely with the predictor, effectively amortizing the cost of running full-training runs. We study JoBS's average regret and devise the optimal budget allocation to minimize regret. JoBS outperforms existing multi-fidelity BO baselines, as well as data and model optimization approaches across diverse LLM tasks under the same optimization budget.

For agents

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

get_harvested_code_for_paper("2602.08351")
get_code_for_paper("2602.08351")
have("2602.08351")

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

curl https://syntology.ai/api/ran/2602.08351.json

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

[![Syntology run record](https://syntology.ai/api/ran/2602.08351.svg)](https://syntology.ai/paper/2602.08351)

Connect an agent — have() is free.