SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2602.13659 · 2026

Zero-Order Optimization for LLM Fine-Tuning via Learnable Direction Sampling

Aleksandr Beznosikov, Andrey Veprikov, † Equal, Valery Parfenov, Grigoriy Evseev, Nikolay Bushkov, Stanislav Moiseev

arXiv · PDF · Open in the Atlas

Code that ran

1 of 1 samples (functions and classes) lifted from this paper's linked repositories executed on a synthesized input in Syntology's jail (1 of them constructed an object rather than computing a result; 0 honoured the contract we drafted, 0 violated it and 1 has no contract verdict; 0 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/2602.13659.json · badge: /api/ran/2602.13659.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
brain-lab-research/zo_ldsd community
Papers with Code archive
1 ran · 0 no recorded run
FunctionStatusWhere it lives
VectorSampler Ran
ran on a synthesized input · constructed an object; no result was computed
brain-lab-research/zo_ldsd/optimizers/opt_utils/vector_sampling.py
pointer only (licence: NONE) · get_code("89cfdb9fbba0013c")
How it ran

The body of this sample is withheld: GitHub detected no licence in this paper's repository, brain-lab-research/zo_ldsd 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 1 samples lifted from this paper's linked repositories, 1 carry a function identity and 0 do not. Of the 1 distinct functions with an identity, 0 are also present in other papers whose code we hold (0 with at least one paper dated earlier, 0 the same month, 0 later) and 1 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.

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

Fine-tuning large pretrained language models (LLMs) is a cornerstone of modern NLP, yet its growing memory demands (driven by backpropagation and large optimizer States) limit deployment in resourceconstrained settings. Zero-order (ZO) methods bypass backpropagation by estimating directional derivatives from forward evaluations, offering substantial memory savings. However, classical ZO estimators suffer from high variance and an adverse dependence on the parameter dimensionality d, which has constrained their use to low-dimensional problems. In this work, we propose a policy-driven ZO framework that treats the sampling distribution over perturbation directions as a learnable policy and updates it to reduce the variance of directional estimates. We develop a practical algorithm implementing this idea and provide a theoretical analysis, showing that learned sampling distributions improve the quality of gradient information and relax the explicit dependence on d in convergence bounds. Empirically, we validate the approach on challenging LLM fine-tuning benchmarks, demonstrating substantially improved performance compared to standard ZO baselines. Our results suggest that adaptive direction sampling is a promising route to make ZO fine-tuning viable at scale. The source code is available at https://github.com/brain-lab-research/zo_ldsd.

For agents

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

get_harvested_code_for_paper("2602.13659")
get_code_for_paper("2602.13659")
have("2602.13659")

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

curl https://syntology.ai/api/ran/2602.13659.json

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

[![Syntology run record](https://syntology.ai/api/ran/2602.13659.svg)](https://syntology.ai/paper/2602.13659)

Connect an agent — have() is free.