Papers › Phi: Preference Hijacking in Multi-modal Large Language Models at Inference Time

Phi: Preference Hijacking in Multi-modal Large Language Models at Inference Time

15 Sep 2025EMNLParXiv:2509.12521added by Syntology

Yifan Lan, Yuanpu Cao, Weitong Zhang, Lu Lin, Jinghui Chen

Title, abstract, authors and date from arXiv's metadata (CC0); this paper is not in the Papers with Code archive (frozen 2025-07-28).

Recently, Multimodal Large Language Models (MLLMs) have gained significant attention across various domains. However, their widespread adoption has also raised serious safety concerns. In this paper, we uncover a new safety risk of MLLMs: the output preference of MLLMs can be arbitrarily manipulated by carefully optimized images. Such attacks often generate contextually relevant yet biased responses that are neither overtly harmful nor unethical, making them difficult to detect. Specifically, we introduce a novel method, Preference Hijacking (Phi), for manipulating the MLLM response preferences using a preference hijacked image. Our method works at inference time and requires no model modifications. Additionally, we introduce a universal hijacking perturbation -- a transferable component that can be embedded into different images to hijack MLLM responses toward any attacker-specified preferences. Experimental results across various tasks demonstrate the effectiveness of our approach. The code for Phi is accessible at https://github.com/Yifan-Lan/Phi.

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 13 code samples harvested from 1 repository linked to this paper: 8 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 1 violated, 2 with no contract checked; 5 where Syntology's instrument failed) · 5 unverified. Where Syntology's instrument failed: 4 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: found in paper text by Syntology: 13 samples from 1 repository, 8 ran (3 with no instrument failure; 5 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 2026-09-25 (UTC): the dates Syntology's graph wrote each sample's verification level, not when the code executed.

Yifan-Lan/Phi found in paper text by SyntologyMITSyntology: 13 samples, 8 ran (5 instrument-failed) · 5 unverified report

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

13 samples harvested: 8 ran (of which 0 constructed an object rather than computing a result; 3 with no instrument failure: 0 honoured, 1 violated, 2 with no contract checked; 5 where Syntology's instrument failed) · 5 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
4ran · our draft was wrong
1ran; the contract check could not run
2ran
5unverified

Licence: 0 of the 13 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 Yifan-Lan/Phi. “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.

Yifan-Lan/Phi found in paper text by Syntology 13 samples, 8 ran (5 instrument-failed) · 5 unverified
convert_state_dict in 13 papers trl/models/sd_utils.py ran · our draft was wrong MIT (permissive) · 8f1ce89b816a0734 · report fixture: signature state_dict:dict/0/?, mapping:dict/0/? · returned dict · level recorded 2026-09-24
convert_to_scalar in 24 papers trl/core.py ran · our draft was wrong MIT (permissive) · c32b636f13ed9ee5 · report fixture: signature stats:dict/0/? · returned dict · level recorded 2026-09-25
get_all_parameters in 7 papers trl/models/utils.py ran MIT (permissive) · 2f5119732e8dbd63 · report fixture: signature sub_module:callable/0/?, recurse:bool/0/bool · returned chain · level recorded 2026-09-25
get_image_data_url in 11 papers GPT_test_score_phi.py ran · our draft was wrong MIT (permissive) · d264a33695ddd017 · report fixture: signature image_file:str/0/str, image_format:str/0/str · returned str · level recorded 2026-09-25
iter_params trl/models/utils.py ran MIT (permissive) · 48685a350888e505 · report fixture: signature module:callable/0/?, recurse:bool/0/bool · returned list · level recorded 2026-09-25
read_csv_data in 6 papers GPT_test_score_phi.py ran · our draft was wrong MIT (permissive) · cac35be5cfe9eec0 · report fixture: signature filepath:str/0/str · returned list · level recorded 2026-09-25
strtobool in 7 papers trl/env_utils.py ran · violated contract MIT (permissive) · f0cba5a1fca2a144 · report fixture: signature val:str/0/str · returned bool · level recorded 2026-09-25
top_k_top_p_filtering in 96 papers
census: 8 implementations from 8 papers, 4 distinct outputs across 2 buckets, one shared input per bucket
trl/core.py ran; the contract check could not run (RAISES) MIT (permissive) · 072cc4ae40638ba1 · report fixture: signature logits:Tensor/2/float, top_k:int/0/int, top_p:float/0/float, filter_value:float/0/float, min_tokens_to_keep:int/0/int · returned Tensor · level recorded 2026-09-25
convert_state_dict_to_diffusers trl/models/sd_utils.py unverified MIT (permissive) · fa59c920caf016e6 · report
flatten_dict in 72 papers trl/core.py unverified MIT (permissive) · a51ae45a1f6c0b5b · report
is_transformers_greater_than in 9 papers trl/import_utils.py unverified MIT (permissive) · 2a9ca02cb6e196ff · report
scheduler_step in 5 papers trl/models/modeling_sd_base.py unverified MIT (permissive) · 31a4e1994101bdd7 · report
setup_chat_format trl/models/utils.py unverified MIT (permissive) · ccaedb0c5f904b0b · report

Results from the paper

The Papers with Code archive ends with its 2025-07-28 snapshot. This paper's arXiv identifier, 2509.12521, was issued in September 2025, after that date, so the archive has no leaderboard rows for it.

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