Papers › Evaluating Deep Unlearning in Large Language Models

Evaluating Deep Unlearning in Large Language Models

19 Oct 2024arXiv:2410.15153archive 2025-07-28

Ruihan Wu, Chhavi Yadav, Russ Salakhutdinov, Kamalika Chaudhuri

Machine unlearning is a key requirement of many data protection regulations such as GDPR. Prior work on unlearning has mostly considered superficial unlearning tasks where a single or a few related pieces of information are required to be removed. However, the task of unlearning a fact is much more challenging in recent large language models (LLMs), because the facts in LLMs can be deduced from each other. In this work, we investigate whether current unlearning methods for LLMs succeed beyond superficial unlearning of facts. Specifically, we formally propose a framework and a definition for deep unlearning facts that are interrelated. We design the metric, recall, to quantify the extent of deep unlearning. To systematically evaluate deep unlearning, we construct a synthetic dataset EDU-RELAT, which consists of a synthetic knowledge base of family relationships and biographies, together with a realistic logical rule set that connects them. We use this dataset to test four unlearning methods in four LLMs at different sizes. Our findings reveal that in the task of deep unlearning only a single fact, they either fail to properly unlearn with high recall, or end up unlearning many other irrelevant facts. Our dataset and code are publicly available at: https://github.com/wrh14/deep_unlearning.

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 12 code samples harvested from 1 repository linked to this paper: 9 ran (of which 0 constructed an object rather than computing a result; 5 with no instrument failure: 0 honoured, 1 violated, 4 with no contract checked; 4 where Syntology's instrument failed) · 3 unverified. Where Syntology's instrument failed: 3 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: 12 samples from 1 repository, 9 ran (5 with no instrument failure; 4 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 between 2026-09-24 and 2026-09-26 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed.

wrh14/deep_unlearning officialmentioned in papermentioned on GitHubpytorchMITSyntology: 12 samples, 9 ran (4 instrument-failed) · 3 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

12 samples harvested: 9 ran (of which 0 constructed an object rather than computing a result; 5 with no instrument failure: 0 honoured, 1 violated, 4 with no contract checked; 4 where Syntology's instrument failed) · 3 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
3ran · our draft was wrong
1ran; the contract check could not run
4ran
3unverified

Licence: 0 of the 12 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 wrh14/deep_unlearning. “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”. For 1 sample that ran, the graph records a signature this page does not show: an argument's recorded type or element kind is a printout of an object in Syntology's harness (such as a mock object, a class or a pandas printout), not the type of an input, so it does not describe the call's shape. “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.

wrh14/deep_unlearning official repository 12 samples, 9 ran (4 instrument-failed) · 3 unverified
add_dataset_index utils.py ran · our draft was wrong MIT (permissive) · 669009a59e41c2c9 · report fixture: returned Mock · level recorded 2026-09-26
check_if_in_deductive_closure utils_metric.py ran MIT (permissive) · dd8e0c53b94c1011 · report fixture: signature unlearn_data_id:int/0/int, minimal_set:set/0/?, edge_list:list/1/tuple, edge_type_list:list/1/str, dc_edge_list:list/1/tuple, dc_edge_type_list:list/1/str, rule_list:list/1/Rule · returned bool · level recorded 2026-09-26
compare in 20 papers task_vector.py ran · violated contract MIT (permissive) · f80682d13e9784c8 · report fixture: signature model1:callable/0/?, model2:callable/0/? · returned bool · level recorded 2026-09-26
convert_raw_data_to_model_format in 10 papers data_module.py ran MIT (permissive) · c148b56239314ed5 · report fixture: signature tokenizer:callable/0/?, max_length:int/0/int, question:str/0/str, answer:str/0/str, model_configs:dict/0/? · returned tuple · level recorded 2026-09-26
custom_data_collator in 6 papers data_module.py ran · our draft was wrong MIT (permissive) · 5985b70afc6ef2b0 · report fixture: signature samples:list/1/tuple · returned tuple · level recorded 2026-09-24
custom_data_collator_npo data_module.py ran · our draft was wrong MIT (permissive) · e566e2214d87e9eb · report fixture: signature samples:list/1/tuple · returned tuple · level recorded 2026-09-26
eval_qa_whp evaluate_util.py ran MIT (permissive) · ae2259c810201f91 · report fixture: signature dataset:list/1/dict, whp_model:callable/0/?, tokenizer:callable/0/?, max_new_tokens:int/0/int, qk:str/0/str, ak:str/0/str, question_start_tag:str/0/str, question_end_tag:str/0/str, answer_tag:str/0/str · returned tuple · level recorded 2026-09-26
get_minimal_nec_unlearn_and_not_included_unlearn utils_metric.py ran MIT (permissive) · dae33bffbca69376 · report fixture: signature unlearn_data_id:int/0/int, edge_list:list/1/tuple, edge_type_list:list/1/str, dc_edge_list:list/1/tuple, dc_edge_type_list:list/1/str, rule_list:list/1/Rule, seed:int/0/int · returned set · level recorded 2026-09-26
get_prec_rec_acc utils_metric.py ran; the contract check could not run (RAISES) MIT (permissive) · b94917ff43e0396e · report fixture: signature minimal_set:set/0/?, unlearn_ind:ndarray/1/int · returned tuple · level recorded 2026-09-26
get_model_identifiers_from_yaml in 12 papers utils.py unverified MIT (permissive) · 41b483c67acbde4d · report
load_model in 479 papers task_vector.py unverified MIT (permissive) · fac5049c9b469cf9 · report
unlearn task_vector.py unverified MIT (permissive) · 137b4007a02b0407 · report

Tasks

Machine Unlearning

Results from the paper archive 2025-07-28

No leaderboard rows for this paper in the archive.

Methods

BASESET

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