Papers › Circuit Breaking: Removing Model Behaviors with Targeted Ablation

Circuit Breaking: Removing Model Behaviors with Targeted Ablation

12 Sep 2023arXiv:2309.05973archive 2025-07-28

Maximilian Li, Xander Davies, Max Nadeau

Language models often exhibit behaviors that improve performance on a pre-training objective but harm performance on downstream tasks. We propose a novel approach to removing undesirable behaviors by ablating a small number of causal pathways between model components, with the intention of disabling the computational circuit responsible for the bad behavior. Given a small dataset of inputs where the model behaves poorly, we learn to ablate a small number of important causal pathways. In the setting of reducing GPT-2 toxic language generation, we find ablating just 12 of the 11.6K causal edges mitigates toxic generation with minimal degradation of performance on other inputs.

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; 9 with no instrument failure: 0 honoured, 0 violated, 9 with no contract checked; 0 where Syntology's instrument failed) · 3 unverified. “Where Syntology's instrument failed” counts failures of Syntology's instrument, not of the code; here there are none.

By repository: official repository: 12 samples from 1 repository, 9 ran (9 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.

xanderdavies/circuit-breaking officialmentioned in paperpytorchSyntology: 12 samples, 9 ran (0 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; 9 with no instrument failure: 0 honoured, 0 violated, 9 with no contract checked; 0 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 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.

9ran
3unverified

Licence: 12 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 xanderdavies/circuit-breaking. “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.

xanderdavies/circuit-breaking official repository 12 samples, 9 ran (0 instrument-failed) · 3 unverified
filter_text toxicity/data_preprocessing.py ran fingerprinted no licence file found · pointer only · bb8e7ced3c0abfa9 · report fixture: signature text:str/0/str · returned str · level recorded 2026-09-24
get_data in 223 papers toxicity/data_preprocessing.py ran no licence file found · pointer only · c889aff403e3ccae · report fixture: signature toxicity_scores:list/1/tuple, num_samples:int/0/int, toxicity_range:list/1/int, num_test_samples:int/0/int, num_buckets:int/0/int · returned tuple · level recorded 2026-09-24
get_mnist_3s_w_bad mnist/data.py ran no licence file found · pointer only · ba9b47ba89029051 · report fixture: signature train_dataset_grouped:TensorDataset/0/?, train_dataset:TensorDataset/0/?, test_dataset_grouped:TensorDataset/0/?, test_dataset:TensorDataset/0/?, bad_behaviors:int/0/int · returned tuple · level recorded 2026-09-24
group_mnist mnist/data.py ran no licence file found · pointer only · 4785260563d2c44a · report fixture: signature train_dataset:TensorDataset/0/?, test_dataset:TensorDataset/0/?, seed:int/0/int · returned tuple · level recorded 2026-09-24
l05_regularization mnist/mlp_model.py ran no licence file found · pointer only · e14402ea7d99d023 · report fixture: signature model:callable/0/?, l05_lambda:float/0/float · returned Tensor · level recorded 2026-09-24
l1_regularization in 6 papers mnist/mlp_model.py ran no licence file found · pointer only · 7a90d9617759fa76 · report fixture: signature model:callable/0/?, l1_lambda:float/0/float · returned Tensor · level recorded 2026-09-24
prepare_fixed_demo toxicity/inference.py ran no licence file found · pointer only · 9f61bbbd76ea7d8a · report fixture: signature tokenizer:callable/0/?, batch_size:int/0/int, demo:str/0/str, device:str/0/str · returned Tensor · level recorded 2026-09-24
retrieve_toxic_filtered_data toxicity/data.py ran no licence file found · pointer only · c5ccb3834251f5f3 · report fixture: signature batch_size:int/0/int, split:str/0/str · returned DataLoader · level recorded 2026-09-24
tokenize_and_concatenate_list toxicity/data.py ran no licence file found · pointer only · a2a7bd73e9052bbd · report fixture: signature text_samples:list/1/str, tokenizer:callable/0/?, seq_len:int/0/int · returned Tensor · level recorded 2026-09-24
epoch in 23 papers mnist/train_eval.py unverified no licence file found · pointer only · 5cefb8a41654c608 · report
infer_batch_with_owt toxicity/inference.py unverified no licence file found · pointer only · 8fbf87f49043e30f · report
retrieve_toxic_data_low_loss toxicity/data.py unverified no licence file found · pointer only · 306013836f145d62 · report

Tasks

Text Generationmodel

Results from the paper archive 2025-07-28

No leaderboard rows for this paper in the archive.

Methods

AdamAttentionAttention DropoutBPECosine AnnealingDense ConnectionsDiscriminative Fine-TuningDropoutGPT-2Layer NormalizationLinear LayerLinear Warmup With Cosine AnnealingMulti-Head AttentionResidual ConnectionSoftmaxWeight Decay

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