SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2604.00536 · 2026

Optimsyn: Influence-Guided Rubrics Optimization for Synthetic Data Generation

Yixin Chen, Junbo Zhao, Jian Wu, Tianxiang Hu, Zuozhu Liu, Ruizhe Chen, Zhiting Fan, Ru Peng, Zenan Huang, Haokai Xu

arXiv · PDF · Open in the Atlas

Code that ran

11 of 16 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, 0 violated it and 1 has no contract verdict; 8 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/2604.00536.json · badge: /api/ran/2604.00536.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
FanZT6/OptimSyn canonical (classifier label)
classifier basis: census
11 ran · 5 no recorded run
FunctionStatusWhere it lives
encode_with_prompt_completion_format Ran
ran; our contract draft was wrong, not the code
FanZT6/OptimSyn/training/influence/get_training_dataset.py
pointer only (licence: NONE) · get_code("7e98a164bbb50ba8")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, FanZT6/OptimSyn 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.

extract_json_from_text Ran
ran; our contract draft was wrong, not the code
FanZT6/OptimSyn/data_synthesis/generate_qa.py
pointer only (licence: NONE) · get_code("4a112a7b36bef292")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, FanZT6/OptimSyn 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.

extract_json_from_text Ran
ran; our contract draft was wrong, not the code
FanZT6/OptimSyn/data_synthesis/generate_qa_eval.py
pointer only (licence: NONE) · get_code("5dd1d503bf65c0c3")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, FanZT6/OptimSyn 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_dataloader Ran
ran on a synthesized input
FanZT6/OptimSyn/training/influence/get_validation_dataset.py
pointer only (licence: NONE) · get_code("2f90db37f34dff6b")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, FanZT6/OptimSyn 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_max_saved_index Ran
ran, honoured the contract we drafted
FanZT6/OptimSyn/training/influence/collect_grad_reps.py
pointer only (licence: NONE) · get_code("ed43dd7816626e87")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, FanZT6/OptimSyn 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
FanZT6/OptimSyn/training/influence/collect_grad_reps.py
pointer only (licence: NONE) · get_code("808c09bce5840bb9")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, FanZT6/OptimSyn 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_jsonl Ran
ran; our contract draft was wrong, not the code
FanZT6/OptimSyn/data_synthesis/generate_qa.py
pointer only (licence: NONE) · get_code("fc7b419016b1e837")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, FanZT6/OptimSyn 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_jsonl Ran
ran; our contract draft was wrong, not the code
FanZT6/OptimSyn/data_synthesis/generate_rubrics.py
pointer only (licence: NONE) · get_code("131f8568b132a083")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, FanZT6/OptimSyn 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_prompt_template Ran
ran; our contract draft was wrong, not the code
FanZT6/OptimSyn/data_synthesis/generate_qa.py
pointer only (licence: NONE) · get_code("870b252c8dbfcaba")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, FanZT6/OptimSyn 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_prompt_template Ran
ran; our contract draft was wrong, not the code
FanZT6/OptimSyn/data_synthesis/generate_qa_eval.py
pointer only (licence: NONE) · get_code("acc9101b33043248")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, FanZT6/OptimSyn 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
FanZT6/OptimSyn/training/influence/get_validation_dataset.py
pointer only (licence: NONE) · get_code("5d4f58c63866080e")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, FanZT6/OptimSyn 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_data_statistics Not yet run
unverified (harvested, no recorded run)
FanZT6/OptimSyn/training/influence/warmup/data_arguments.py
pointer only (licence: NONE) · get_code("fc5b455108e53a6e")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, FanZT6/OptimSyn 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.

get_encode_function Not yet run
unverified (harvested, no recorded run)
FanZT6/OptimSyn/training/influence/get_training_dataset.py
pointer only (licence: NONE) · get_code("f45c88614f6f7fd4")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, FanZT6/OptimSyn 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.

get_training_dataset Not yet run
unverified (harvested, no recorded run)
FanZT6/OptimSyn/training/influence/get_training_dataset.py
pointer only (licence: NONE) · get_code("b410f4116bd363eb")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, FanZT6/OptimSyn 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.

load_jsonl Not yet run
unverified (harvested, no recorded run)
FanZT6/OptimSyn/data_synthesis/generate_qa_eval.py
pointer only (licence: NONE) · get_code("761da305e0f257c4")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, FanZT6/OptimSyn 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)
FanZT6/OptimSyn/training/influence/warmup/data_arguments.py
pointer only (licence: NONE) · get_code("12b6473bde25b1dd")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, FanZT6/OptimSyn 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 16 samples lifted from this paper's linked repositories, 16 carry a function identity and 0 do not. Of the 16 distinct functions with an identity, 6 are also present in other papers whose code we hold (6 with at least one paper dated earlier, 0 the same month, 0 later) and 10 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
encode_with_prompt_completion_format 9 other papers
9 dated earlier
2311.05657 (2023-11)
Agent Lumos: Unified and Modular Training for Open-Source Language Agents
2312.07540 (2023-12)
diff History for Neural Language Agents
2401.08565 (2024-01)
Tuning Language Models by Proxy
and 6 more not named here
get_max_saved_index 4 other papers
4 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
2411.12882 (2024-11)
ProSec: Fortifying Code LLMs with Proactive Security Alignment
and 1 more not named here
get_output 4 other papers
4 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
2411.12882 (2024-11)
ProSec: Fortifying Code LLMs with Proactive Security Alignment
and 1 more not named here
get_data_statistics 3 other papers
3 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
2602.08351 (2026-02)
The Chicken and Egg Dilemma: Co-optimizing Data and Model Configurations for LLMs
none_or_str 3 other papers
3 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
2602.08351 (2026-02)
The Chicken and Egg Dilemma: Co-optimizing Data and Model Configurations for LLMs
tokenize 3 other papers
3 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
2602.08351 (2026-02)
The Chicken and Egg Dilemma: Co-optimizing Data and Model Configurations for LLMs

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

Large language models (LLMs) achieve strong downstream performance largely due to abundant supervised fine-tuning (SFT) data that imparts problem-solving capabilities. However, as applications expand, high-quality SFT data in knowledge-intensive domains (e.g., humanities and social sciences, medicine, law, finance) is exceedingly scarce: expert curation is costly, privacy constraints are strict, and label consistency is hard to guarantee. Recent work turns to synthetic data, typically prompting a generator over domain documents and filtering with handcrafted rubrics. Yet, rubric design is expert-dependent and rarely transfers across domains; moreover, prevalent heuristic optimization follows a brittle loop (write rubric → synthesize → train → inspect → guess tweaks) that lacks reliable, quantitative feedback about a rubric's true contribution to downstream performance. We argue for assessing synthetic data quality through its training utility on the target model, using this feedback to guide data generation. Inspired by classic influence estimations, we repurpose an optimizer-aware estimator that uses gradient information to quantify each synthetic sample's contribution to the objective of a given target model on specific tasks. Our analysis reveals a gap: although synthetic and real samples may be close in embedding space, their influence on learning can differ substantially. Building on this insight, we propose an optimization-based synthetic data framework that adapts rubrics with targetmodel feedback. Instead of manually engineering domain rubrics, we supply lightweight guiding text and delegate rubric generation to a rubric-specialized model conditioned on the task; crucially, we employ influence score as reward and optimize the rubric generator with reinforcement learning. Empirically, the framework yields consistent gains across domains (HSS and health), target models (e.g., Qwen and Llama families), and data generators, demonstrating broad generalization and engineering portability without task-specific tuning. Code is available at https://github.com/FanZT6/OptimSyn.

For agents

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

get_harvested_code_for_paper("2604.00536")
get_code_for_paper("2604.00536")
have("2604.00536")

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

curl https://syntology.ai/api/ran/2604.00536.json

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

[![Syntology run record](https://syntology.ai/api/ran/2604.00536.svg)](https://syntology.ai/paper/2604.00536)

Connect an agent — have() is free.