Papers › DPZero: Private Fine-Tuning of Language Models without Backpropagation

DPZero: Private Fine-Tuning of Language Models without Backpropagation

14 Oct 2023arXiv:2310.09639archive 2025-07-28

Liang Zhang, Bingcong Li, Kiran Koshy Thekumparampil, Sewoong Oh, Niao He

The widespread practice of fine-tuning large language models (LLMs) on domain-specific data faces two major challenges in memory and privacy. First, as the size of LLMs continues to grow, the memory demands of gradient-based training methods via backpropagation become prohibitively high. Second, given the tendency of LLMs to memorize training data, it is important to protect potentially sensitive information in the fine-tuning data from being regurgitated. Zeroth-order methods, which rely solely on forward passes, substantially reduce memory consumption during training. However, directly combining them with standard differentially private gradient descent suffers more as model size grows. To bridge this gap, we introduce DPZero, a novel private zeroth-order algorithm with nearly dimension-independent rates. The memory efficiency of DPZero is demonstrated in privately fine-tuning RoBERTa and OPT on several downstream tasks. Our code is available at https://github.com/Liang137/DPZero.

PaperPDFCodeCode Syntology ran

In Syntology Open this paper in Syntology's Atlas, the map of the papers in Syntology's graph and their citations.

For agents, Syntology's MCP tool lists every function and class Syntology harvested from this paper and whether it ran (how to connect): get_harvested_code_for_paper(arxiv_id="2310.09639")

Code

Syntology Ran 7 of 17 code samples harvested from 1 repository linked to this paper; 10 have no recorded run. Of those that ran: 2 ran · our draft was wrong; 5 ran with no contract checked.

By repository: official repository: 17 samples from 1 repository, 7 ran. The run record, sample by sample. “Ran” means executed on a synthesized input, not that the code is correct or reproduces the paper.

liang137/dpzero officialmentioned in papermentioned on GitHubpytorchMIT report

Repository list and official/mentioned flags are the archive's, frozen 2025-07-28. Reachability, where shown, is from one Syntology probe window (2026-09-16 to 2026-09-18); repositories not probed show nothing. GitHub stars are not tracked.

Code Syntology ran Syntology

17 samples harvested; 7 ran; 0 honoured the contract we drafted; 10 have no recorded run. Read from Syntology's graph 2026-09-24; that is when this build read the record, not when the samples ran.

2ran · our draft was wrong
5ran
10unverified

Licence: 0 of the 17 samples are pointer only, meaning Syntology does not serve that copy's text. This page shows no code text for any sample; each one links to its file in the repository.

Harvested from Liang137/DPZero. “Ran” means the sample executed on a synthesized input. It does not mean the output is correct, and nothing here reproduces the paper's results. “Honoured” and “violated” refer to a contract Syntology drafted from the code itself; “our draft was wrong” and “fixture could not drive it” are failures of Syntology's instrument, not of the code.

Each sample ends with its code_sha256, Syntology's identity for that exact code. An agent fetches the stored sample with Syntology's MCP tool get_code(code_sha256="…") (how to connect); click an identity to copy that call.

Repository labels, per sample. official repository: The archive marks this repository official for the paper. named in the paper: The archive records that the paper mentions this repository; it is not marked official. community (archive-listed): In the archive's code links for this paper, not marked official and not recorded as mentioned in the paper. found in paper text by Syntology: Syntology found this repository in the paper's own text; whether it is the authors' implementation is not asserted. community: Not in the archive's code links for this paper; a community repository Syntology harvested. Samples from a repository marked official are listed first. Licence labels name the repository's licence as recorded at harvest. “Pointer only” means Syntology does not serve that copy's text, for one of four reasons: no licence file was found; the licence was not identified; the licence is recorded as permissive but that copy's record is not marked cleared; or the licence is outside the permissive list Syntology serves text under (MIT, Apache-2.0, BSD and similar). Some licences outside that list permit redistribution, such as WTFPL, and GPL-3.0 under its conditions; they are simply not on the list. Hover a licence label for the reason. File links open the file on GitHub at the default branch, which may have changed since the harvest.

encode_prompt Liang137/DPZero/opt/src/utils.py official repository ran MIT (permissive) · 7c5d9ba063f6642c · report
find_module Liang137/DPZero/opt/src/lora.py official repository ran · our draft was wrong MIT (permissive) · b5adfd6ac99a1b18 · report
forward_wrap_with_option_len_dpzero Liang137/DPZero/opt/src/utils.py official repository ran MIT (permissive) · da6c8809e7b9fce0 · report
input_example_to_string Liang137/DPZero/roberta/src/dataset.py official repository ran MIT (permissive) · e768e25ca95f70cd · report
input_example_to_tuple Liang137/DPZero/roberta/src/dataset.py official repository ran MIT (permissive) · fbef70dc7c700501 · report
normalize_answer Liang137/DPZero/opt/src/metrics.py official repository ran · our draft was wrong fingerprinted MIT (permissive) · e7e75981cb464788 · report
tokenize_multipart_input Liang137/DPZero/roberta/src/dataset.py official repository ran MIT (permissive) · 9a62fa939fc563a4 · report
AsymmetricLstsqKernelSolver liang137/dpzero/roberta/src/kernel_solvers.py official repository unverified MIT (permissive) · 6819ef75ae988d6e · report
BaseKernelSolver liang137/dpzero/roberta/src/kernel_solvers.py official repository unverified MIT (permissive) · e4ad3f4b56f8bf3e · report
LstsqKernelSolver liang137/dpzero/roberta/src/kernel_solvers.py official repository unverified MIT (permissive) · 4d101a4f2733809f · report
attn_forward_hook Liang137/DPZero/opt/src/prefix.py official repository unverified MIT (permissive) · bdb460cb247501dc · report
calculate_metric Liang137/DPZero/opt/src/metrics.py official repository unverified MIT (permissive) · 84c658f2c9252878 · report
f1 Liang137/DPZero/opt/src/metrics.py official repository unverified MIT (permissive) · b989d4bce26f77ce · report
get_task Liang137/DPZero/opt/src/tasks.py official repository unverified MIT (permissive) · f7c384f1e681ea7a · report
prepare_inputs_for_generation Liang137/DPZero/opt/src/prefix.py official repository unverified MIT (permissive) · 84920a8095dca4fa · report
tensor_all_gather Liang137/DPZero/roberta/src/linearhead_trainer.py official repository unverified MIT (permissive) · c239cb1add74b809 · report
varsize_tensor_all_gather Liang137/DPZero/roberta/src/linearhead_trainer.py official repository unverified MIT (permissive) · 734d0491ecbc7841 · report

Results from the paper archive 2025-07-28

No leaderboard rows for this paper in the archive.

Methods

AdamAttentionAttention DropoutBERTDense ConnectionsDropoutLayer NormalizationLinear LayerLinear Warmup With Linear DecayMulti-Head AttentionOPTResidual ConnectionRoBERTaSoftmaxWeight DecayWordPiece

Report a problem or propose a change · a person checks every report against the paper or source before anything changes; decisions are listed on /corrections