Papers › Differential Privacy, Linguistic Fairness, and Training Data Influence: Impossibility...

Differential Privacy, Linguistic Fairness, and Training Data Influence: Impossibility and Possibility Theorems for Multilingual Language Models

17 Aug 2023arXiv:2308.08774archive 2025-07-28

Phillip Rust, Anders Søgaard

Language models such as mBERT, XLM-R, and BLOOM aim to achieve multilingual generalization or compression to facilitate transfer to a large number of (potentially unseen) languages. However, these models should ideally also be private, linguistically fair, and transparent, by relating their predictions to training data. Can these requirements be simultaneously satisfied? We show that multilingual compression and linguistic fairness are compatible with differential privacy, but that differential privacy is at odds with training data influence sparsity, an objective for transparency. We further present a series of experiments on two common NLP tasks and evaluate multilingual compression and training data influence sparsity under different privacy guarantees, exploring these trade-offs in more detail. Our results suggest that we need to develop ways to jointly optimize for these objectives in order to find practical trade-offs.

PaperPDFCode Syntology ran

In Syntology View this paper on Syntology, its page in Syntology's graph. That page adds which of its functions are also present in other papers' code Syntology holds, with dates, and its dated run record as JSON and a README badge. It also lists the repositories linked to the paper, the abstract and the calls for agents.

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

For agents, on Syntology's MCP service (how to connect):

Code

Syntology 23 code samples harvested from 3 repositories linked to this paper: 14 ran (of which 0 constructed an object rather than computing a result; 14 with no instrument failure: 0 honoured, 0 violated, 14 with no contract checked; 0 where Syntology's instrument failed) · 9 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code; here there are none.

By repository: found in paper text by Syntology: 23 samples from 3 repositories, 14 ran (14 with no instrument failure; 0 where Syntology's instrument failed). The run record, sample by sample. “Ran” means executed on a synthesized input, not that the code is correct or reproduces the paper.

No sample is linked to this paper from 1 of its repositories that Syntology harvested for other papers:

Levels written on 2026-09-24 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed.

huggingface/transformers found in paper text by SyntologySyntology: no sample linked to this paper (harvested for another paper). Syntology's graph links these samples to Attention Is All You Need (arXiv:1706.03762; that paper's own run record: 610 ran (of which 293 constructed an object rather than computing a result; 529 with no instrument failure: 45 honoured, 15 violated, 469 with no contract checked; 81 where Syntology's instrument failed) · 336 unverified), HuggingFace's Transformers: State-of-the-art Natural Language Processing (arXiv:1910.03771; that paper's own run record: 3 ran (of which 0 constructed an object rather than computing a result; 0 with no instrument failure: 0 honoured, 0 violated, 0 with no contract checked; 3 where Syntology's instrument failed) · 0 unverified), DeBERTa: Decoding-enhanced BERT with Disentangled Attention (arXiv:2006.03654; that paper's own run record: 4 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 2 honoured, 1 violated, 0 with no contract checked; 1 where Syntology's instrument failed) · 9 unverified) and 1 more (every one in this page's JSON twin); nothing checks that they implement this paper's method. report
LBeaudoux/tatoebatools found in paper text by SyntologyMITSyntology: 2 samples, 2 ran (0 instrument-failed) · 0 unverified report
xplip/multilingual-lm-objectives found in paper text by SyntologyMITSyntology: 11 samples, 7 ran (0 instrument-failed) · 4 unverified report
lxuechen/private-transformers found in paper text by SyntologyApache-2.0Syntology: 10 samples, 5 ran (0 instrument-failed) · 5 unverified 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.

A paper named beside a repository with no sample linked to this paper is shown with that paper's own run record, not this paper's: “ran” means executed on a synthesized input, not that the code is correct, and “where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Papers are listed in arXiv-id order, at most three per repository.

Code Syntology ran Syntology

23 samples harvested: 14 ran (of which 0 constructed an object rather than computing a result; 14 with no instrument failure: 0 honoured, 0 violated, 14 with no contract checked; 0 where Syntology's instrument failed) · 9 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Syntology's record for this page has not changed since , the first build that kept a record date for it; when this build read Syntology's graph is in the build record. That date is not when the samples ran.

14ran
9unverified

Licence: 0 of the 23 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 3 repositories linked to this paper, official or community; the list below is grouped by repository, each group headed by its name and what it is to the paper. “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. A line beginning “fixture:” under a sample that ran gives the arguments Syntology's fixture called the sample with, each as name:type/rank/element kind (? where none was recorded), and the type of what the fixture's run returned, exactly as the harness recorded them; the arguments and the returned type were recorded by two separate runs of the same fixture. No output value is shown, and a returned type says nothing about whether the output was right. “Level recorded” gives the UTC date when Syntology's graph wrote the sample's verification level, not when the code executed; where the level was written before Syntology's graph kept that date, the row says “level recorded, date not recorded”. “Honoured” and “violated” refer to a contract Syntology drafted from the code itself. “Our draft was wrong” and “the contract check could not run” are failures of Syntology's instrument, not of the code: the sample ran on Syntology's synthesized input, and then the drafted contract did not describe it, or the check itself raised an error (RAISES), lacked a dependency (DEP_MISSING) or ran out of time (TIMEOUT); hover the label for which. A metamorphic tier, where shown, is evidence of the output's shape, not that the sample implements the method.

Each sample ends with its code_sha256. A sample's code_sha256 is Syntology's identity for that exact code. It is printed in full (16 hexadecimal characters for every sample in this build), and get_code(code_sha256="…") takes it exactly as printed. An agent calls it on Syntology's MCP service (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 are grouped by repository: a repository marked official first, then by how directly the repository is tied to the paper, then in the archive's order; never by how many ran. 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 six 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; the licence is outside the permissive list Syntology serves text under (MIT, Apache-2.0, BSD and similar); this paper's copy was not located, so its licence is not recorded; or this paper's copy is cleared, but the stored copy of the same code that Syntology serves is not. 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.

A function or class name links to its page when the name occurs in the harvested code of at least 5 papers; the name is a string, not an identity. A census tag appears when Syntology's census ran implementations with that name on one shared input per bucket: it gives how many were compared, from how many papers, and how many distinct outputs they produced, and names this sample's bucket when this sample was one of them.

LBeaudoux/tatoebatools found in paper text by Syntology 2 samples, 2 ran (0 instrument-failed) · 0 unverified
compare_csv tatoebatools/difference.py ran MIT (permissive) · 6d26d45c8e798207 · report fixture: signature csv_left:PosixPath/0/?, csv_right:PosixPath/0/?, delimiter:str/0/str, quoting:int/0/int, index_col_keys:NoneType/0/?, verbose:bool/0/bool · returned dict · level recorded 2026-09-24
reset_pos tatoebatools/datafile.py ran MIT (permissive) · fb96d82eef598129 · report fixture: signature func:callable/0/? · returned function · level recorded 2026-09-24
xplip/multilingual-lm-objectives found in paper text by Syntology 11 samples, 7 ran (0 instrument-failed) · 4 unverified
compute_loss in 97 papers
census: 11 implementations from 11 papers, 9 distinct outputs across 7 buckets, one shared input per bucket
finetuning/pos/tracin_pos.py ran MIT (permissive) · 88380f81794a2565 · report fixture: signature outputs:Tensor/2/float, targets:Tensor/1/int · returned Tensor · level recorded 2026-09-24
convert_examples_to_features in 134 papers finetuning/pos/pos_dataset.py ran MIT (permissive) · c7dd8be65b126cb4 · report fixture: signature examples:list/1/InputExample, label_list:list/1/str, max_seq_length:int/0/int, tokenizer:MockTokenizer/0/?, cls_token_at_end:bool/0/bool, cls_token:str/0/str, cls_token_segment_id:int/0/int, sep_token:str/0/str, sep_token_extra:bool/0/bool, pad_on_left:bool/0/bool, pad_token:int/0/int, pad_token_segment_id:int/0/int, pad_token_label_id:int/0/int, sequence_a_segment_id:int/0/int, mask_padding_with_zero:bool/0/bool · returned list · level recorded 2026-09-24
evaluate_retrieval in 7 papers evaluation/evaluation_utils.py ran MIT (permissive) · c11453363b10dd6a · report fixture: signature device:device/0/?, vectors:list/1/Tensor · returned tuple · level recorded 2026-09-24
get_distances in 12 papers evaluation/evaluation_utils.py ran fingerprinted MIT (permissive) · f613227e9f1c6c9c · report fixture: signature X:Tensor/2/float, Y:Tensor/2/float · returned Tensor · level recorded 2026-09-24
get_grad_z finetuning/pos/tracin_pos.py ran MIT (permissive) · 03f31dbf34c04281 · report fixture: signature inputs_z:dict/0/?, model:callable/0/?, last_n_encoder_layers:int/0/int · returned list · level recorded 2026-09-24
map_a_special_embeddings_into_b utils/dp_training_utils.py ran MIT (permissive) · b879433be37a1f1f · report fixture: signature a_embeddings:callable/0/?, a_tokenizer:StandInTokenizer/0/?, b_embeddings:callable/0/?, b_tokenizer:StandInTokenizer/0/? · returned StandInEmbedding · level recorded 2026-09-24
remove_padding in 5 papers evaluation/evaluation_utils.py ran MIT (permissive) · 41c6daed8df5aa21 · report fixture: signature tensor:Tensor/3/float, padding_mask:Tensor/2/bool · returned list · level recorded 2026-09-24
get_file in 18 papers finetuning/pos/pos_dataset.py unverified MIT (permissive) · 9c85503addee47df · report
load_checkpoint in 277 papers finetuning/pos/tracin_pos.py unverified MIT (permissive) · eedeab21f434a49e · report
load_checkpoint in 277 papers finetuning/xnli/tracin_xnli.py unverified MIT (permissive) · 96cf21cd04f4f381 · report
read_examples_from_file in 11 papers finetuning/pos/pos_dataset.py unverified MIT (permissive) · 59e3a2396c98aeee · report
lxuechen/private-transformers found in paper text by Syntology 10 samples, 5 ran (0 instrument-failed) · 5 unverified
convert_gpt2_attention_to_lora private_transformers/lora_utils.py ran Apache-2.0 (permissive) · 8eda8261353a36a2 · report fixture: signature model:callable/0/?, lora_r:int/0/int, lora_alpha:float/0/float, lora_dropout:float/0/float · returned GPT2LMHeadModel · level recorded 2026-09-24
get_privacy_spent private_transformers/accounting/rdp_accounting.py ran fingerprinted Apache-2.0 (permissive) · 5429ceb040535efa · report fixture: signature orders:ndarray/1/float, rdp:ndarray/1/float, delta:float/0/float · returned tuple · level recorded 2026-09-24
requires_grad in 5 papers private_transformers/autograd_grad_sample.py ran Apache-2.0 (permissive) · 5bfe971f0927820a · report fixture: signature module:callable/0/?, recurse:bool/0/bool · returned bool · level recorded 2026-09-24
sum_over_all_but_batch_and_last_n private_transformers/supported_layers_grad_samplers.py ran fingerprinted Apache-2.0 (permissive) · d62ca268adb92593 · report fixture: signature tensor:Tensor/4/float, n_dims:int/0/int · returned Tensor · level recorded 2026-09-24
unfold2d private_transformers/supported_layers_grad_samplers.py ran Apache-2.0 (permissive) · 088a357156c09c02 · report fixture: signature input:Tensor/4/float, kernel_size:tuple/1/int, padding:tuple/1/int, stride:tuple/1/int, dilation:tuple/1/int · returned Tensor · level recorded 2026-09-24
add_hooks private_transformers/autograd_grad_sample.py unverified Apache-2.0 (permissive) · e7bbbbdcc4595ebd · report
compute_rdp in 5 papers private_transformers/accounting/rdp_accounting.py unverified Apache-2.0 (permissive) · 9f57ed8d21d5b2de · report
swap_bert_model_forward private_transformers/transformers_support.py unverified Apache-2.0 (permissive) · 05fcbe0865ca943a · report
swap_gpt2_model_forward private_transformers/transformers_support.py unverified Apache-2.0 (permissive) · e6bec88416415966 · report
swap_openai_gpt_model_forward private_transformers/transformers_support.py unverified Apache-2.0 (permissive) · d6ada14227fbf040 · report

Tasks

FairnessXLM-R

Results from the paper archive 2025-07-28

No leaderboard rows for this paper in the archive.

Methods

BLOOMXLM-RmBERT

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