Papers › Tamper-Resistant Safeguards for Open-Weight LLMs

Tamper-Resistant Safeguards for Open-Weight LLMs

1 Aug 2024arXiv:2408.00761archive 2025-07-28

Rishub Tamirisa, Bhrugu Bharathi, Long Phan, Andy Zhou, Alice Gatti, Tarun Suresh, Maxwell Lin, Justin Wang, Rowan Wang, Ron Arel, Andy Zou, Dawn Song, Bo Li, Dan Hendrycks, Mantas Mazeika

Rapid advances in the capabilities of large language models (LLMs) have raised widespread concerns regarding their potential for malicious use. Open-weight LLMs present unique challenges, as existing safeguards lack robustness to tampering attacks that modify model weights. For example, recent works have demonstrated that refusal and unlearning safeguards can be trivially removed with a few steps of fine-tuning. These vulnerabilities necessitate new approaches for enabling the safe release of open-weight LLMs. We develop a method, called TAR, for building tamper-resistant safeguards into open-weight LLMs such that adversaries cannot remove the safeguards even after thousands of steps of fine-tuning. In extensive evaluations and red teaming analyses, we find that our method greatly improves tamper-resistance while preserving benign capabilities. Our results demonstrate that tamper-resistance is a tractable problem, opening up a promising new avenue to improve the safety and security of open-weight LLMs.

PaperPDFCodeCode 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 16 code samples harvested from 2 repositories linked to this paper: 9 ran (of which 0 constructed an object rather than computing a result; 7 with no instrument failure: 0 honoured, 1 violated, 6 with no contract checked; 2 where Syntology's instrument failed) · 7 unverified. Where Syntology's instrument failed: 1 where our contract draft was wrong; 1 where the contract check could not run; failures of Syntology's instrument, not of the code.

By repository: official repository: 14 samples from 1 repository, 9 ran (7 with no instrument failure; 2 where Syntology's instrument failed); community (archive-listed): 2 samples from 1 repository, 0 ran (0 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.

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

rishub-tamirisa/tamper-resistance officialmentioned in papermentioned on GitHubpytorchMITSyntology: 14 samples, 9 ran (2 instrument-failed) · 5 unverified report
boyiwei/tar-reproduce mentioned on GitHubpytorchMITSyntology: 2 samples, 0 ran · 2 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.

Code Syntology ran Syntology

16 samples harvested: 9 ran (of which 0 constructed an object rather than computing a result; 7 with no instrument failure: 0 honoured, 1 violated, 6 with no contract checked; 2 where Syntology's instrument failed) · 7 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code. Syntology's record for this page last changed ; when this build read Syntology's graph is in the build record. That date is not when the samples ran.

1ran · violated contract
1ran · our draft was wrong
1ran; the contract check could not run
6ran
7unverified

Licence: 0 of the 16 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 2 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.

rishub-tamirisa/tamper-resistance official repository 14 samples, 9 ran (2 instrument-failed) · 5 unverified
get_adam red_teaming/optimizers.py ran MIT (permissive) · 2d4ef3518e6098f8 · report fixture: signature model:callable/0/?, learning_rate:float/0/float, momentum:float/0/float, warmup_steps:int/0/int · returned Adam · level recorded 2026-09-24
get_sgd_with_momentum red_teaming/optimizers.py ran MIT (permissive) · 4af747db2262038a · report fixture: signature model:callable/0/?, learning_rate:float/0/float, momentum:float/0/float, warmup_steps:NoneType/0/? · returned SGD · level recorded 2026-09-24
get_sgd_with_nesterov_momentum red_teaming/optimizers.py ran MIT (permissive) · 1f164099e83b0e6f · report fixture: signature model:callable/0/?, learning_rate:float/0/float, momentum:float/0/float, warmup_steps:NoneType/0/? · returned SGD · level recorded 2026-09-24
lambda_fn tar.py ran; the contract check could not run (RAISES) MIT (permissive) · d0ee74a4bdf0b184 · report fixture: signature module:callable/0/? · returned bool · level recorded 2026-09-24
log_1_minus_p_loss modules/objectives.py ran MIT (permissive) · 5cb0ed367e769a24 · report fixture: signature _logits:Tensor/3/float, _labels:Tensor/2/int, vocab_size:int/0/int, threshold:float/0/float · returned Tensor · level recorded 2026-09-24
log_p_loss modules/objectives.py ran MIT (permissive) · d2fee7baeb212dfd · report fixture: signature logits:Tensor/3/float, labels:Tensor/2/int, vocab_size:int/0/int · returned Tensor · level recorded 2026-09-24
max_entropy_loss modules/objectives.py ran · our draft was wrong fingerprinted MIT (permissive) · ca477c3a5169e14c · report fixture: signature logits:Tensor/2/float · returned Tensor · level recorded 2026-09-24
return_coin_flip_batch_selection modules/utils.py ran · violated contract MIT (permissive) · d2d6bebf37fdf81c · report fixture: signature prob_of_selecting_retain_batch:float/0/float · returned Tensor · level recorded 2026-09-24
return_step_based_batch_selection modules/utils.py ran MIT (permissive) · 65d8919abbae12b8 · report fixture: signature current_step:int/0/int, max_steps:int/0/int, prop_steps_for_batch_selection:float/0/float · returned bool · level recorded 2026-09-24
get_bio_pilecamel_forget_with_heldout_dataloaders modules/dataloaders.py unverified MIT (permissive) · 9d75b6ae75b4e40f · report
get_camel_ai_datasets modules/dataloaders.py unverified MIT (permissive) · 1a17a7e1f54cb58c · report
get_distributed_random_number modules/utils.py unverified MIT (permissive) · d8079961a9ada7e2 · report
get_pile_bio_retain_forget_heldout_datasets modules/dataloaders.py unverified MIT (permissive) · 5051961408d3c896 · report
lambda_fn modules/baselines.py unverified MIT (permissive) · 84bd023f74905cd6 · report
boyiwei/tar-reproduce community (archive-listed) 2 samples, 0 ran · 2 unverified
get_bio_pilecamel_forget_with_heldout_dataloaders modules/dataloaders.py unverified MIT (permissive) · 2066015c29979283 · report
get_pile_bio_retain_forget_heldout_datasets modules/dataloaders.py unverified MIT (permissive) · ba92aaf6d0cfbfc9 · report

Tasks

Red TeamingTAR

Results from the paper archive 2025-07-28

No leaderboard rows for this paper in the archive.

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