Papers › Causality Analysis for Evaluating the Security of Large Language Models

Causality Analysis for Evaluating the Security of Large Language Models

13 Dec 2023arXiv:2312.07876archive 2025-07-28

Wei Zhao, Zhe Li, Jun Sun

Large Language Models (LLMs) such as GPT and Llama2 are increasingly adopted in many safety-critical applications. Their security is thus essential. Even with considerable efforts spent on reinforcement learning from human feedback (RLHF), recent studies have shown that LLMs are still subject to attacks such as adversarial perturbation and Trojan attacks. Further research is thus needed to evaluate their security and/or understand the lack of it. In this work, we propose a framework for conducting light-weight causality-analysis of LLMs at the token, layer, and neuron level. We applied our framework to open-source LLMs such as Llama2 and Vicuna and had multiple interesting discoveries. Based on a layer-level causality analysis, we show that RLHF has the effect of overfitting a model to harmful prompts. It implies that such security can be easily overcome by `unusual' harmful prompts. As evidence, we propose an adversarial perturbation method that achieves 100% attack success rate on the red-teaming tasks of the Trojan Detection Competition 2023. Furthermore, we show the existence of one mysterious neuron in both Llama2 and Vicuna that has an unreasonably high causal effect on the output. While we are uncertain on why such a neuron exists, we show that it is possible to conduct a ``Trojan'' attack targeting that particular neuron to completely cripple the LLM, i.e., we can generate transferable suffixes to prompts that frequently make the LLM produce meaningless responses.

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 9 code samples harvested from 1 repository linked to this paper: 7 ran (of which 0 constructed an object rather than computing a result; 4 with no instrument failure: 0 honoured, 0 violated, 4 with no contract checked; 3 where Syntology's instrument failed) · 2 unverified. Where Syntology's instrument failed: 3 where our contract draft was wrong; failures of Syntology's instrument, not of the code.

By repository: official repository: 9 samples from 1 repository, 7 ran (4 with no instrument failure; 3 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 2026-09-22 and 2026-09-25 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed.

casperllm/casper officialmentioned in paperpytorchSyntology: 9 samples, 7 ran (3 instrument-failed) · 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

9 samples harvested: 7 ran (of which 0 constructed an object rather than computing a result; 4 with no instrument failure: 0 honoured, 0 violated, 4 with no contract checked; 3 where Syntology's instrument failed) · 2 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.

3ran · our draft was wrong
4ran
2unverified

Licence: 9 of the 9 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 casperllm/casper. “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.

casperllm/casper official repository 9 samples, 7 ran (3 instrument-failed) · 2 unverified
decode_tokens in 8 papers utils/modelUtils.py ran no licence file found · pointer only · 9532c9c8f3eec45b · report fixture: signature tokenizer:MockTokenizer/0/?, token_array:ndarray/2/int · returned list · level recorded 2026-09-25
generate_input in 6 papers utils/utils.py ran no licence file found · pointer only · 6f465ce5300a3f4d · report fixture: signature conv_template:ConvTemplate/0/?, prompt:str/0/str, adv_suffix:NoneType/0/? · returned str · level recorded 2026-09-24
generate_outputs utils/modelUtils.py ran no licence file found · pointer only · 333d6c519f821492 · report fixture: signature current_test_cases:list/1/str, mt:MockModelTokenizerWrapper/0/?, device:str/0/str, batch_size:int/0/int, max_new_tokens:int/0/int, verbose:bool/0/bool · returned list · level recorded 2026-09-24
hierarchical_subsequence in 47 papers casper/nethook.py ran no licence file found · pointer only · 920b394e7ad80c53 · report fixture: signature sequential:callable/0/?, first:NoneType/0/?, last:NoneType/0/?, after:NoneType/0/?, upto:NoneType/0/?, share_weights:bool/0/bool, depth:int/0/int · returned Sequential · level recorded 2026-09-24
make_inputs in 7 papers utils/modelUtils.py ran · our draft was wrong no licence file found · pointer only · 96b3625b39d4a804 · report fixture: signature tokenizer:MockTokenizer/0/?, prompts:list/1/str, device:str/0/str · returned dict · level recorded 2026-09-24
recursive_copy in 51 papers casper/nethook.py ran · our draft was wrong no licence file found · pointer only · 70f6ab8bde55420e · report fixture: signature x:dict/0/?, clone:bool/0/bool, detach:bool/0/bool, retain_grad:bool/0/bool · returned dict · level recorded 2026-09-22
subsequence in 51 papers casper/nethook.py ran · our draft was wrong no licence file found · pointer only · 440ff98c2b1ae1aa · report fixture: signature sequential:callable/0/?, first_layer:NoneType/0/?, last_layer:NoneType/0/?, after_layer:NoneType/0/?, upto_layer:NoneType/0/?, single_layer:NoneType/0/?, share_weights:bool/0/bool · returned Sequential · level recorded 2026-09-24
trace_with_patch_layer analysis/layer.py unverified no licence file found · pointer only · 204dfaf0e991622b · report
trace_with_patch_neuron analysis/neuron.py unverified no licence file found · pointer only · ee1faeebf5c90755 · report

Tasks

Red Teaming

Results from the paper archive 2025-07-28

No leaderboard rows for this paper in the archive.

Methods

AdamAttentionAttention DropoutBPECosine AnnealingDense ConnectionsDiscriminative Fine-TuningDropoutGPTLayer 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