Paper Citation Record · LEDGER

Image Hijacks: Adversarial Images can Control Generative Models at Runtime

As of 8 August 2026, this Paper Citation Record lists 0 of 0 outbound references and 31 inbound Pith citation observations for arXiv:2309.00236.

A citation records a reference. It does not transfer a finding from one paper to another.

pith.paper-citation-record.v1
2309.00236 v4

Coverage vector

measured 0 of 0 reference resolution

Typed states for the displayed outbound observations.

Source: paper_references, paper_reference_links

measured 31 of 31 standing notices

One-hop event checks from named stored sources.

Source: scholarly_work_events, retraction_status_cache, observed 2026-08-08T06:32:00.761636+00:00

measured 31 of 31 inbound itemization

Pith citing papers itemized under the disclosed page cap.

Source: paper_references, paper_reference_links, observed 2026-08-07T13:35:49.779331Z

measured 1 of 1 external citation measurements

A source-named dated measurement, never combined with another source.

Source: pith, observed 2026-08-05T02:28:24.338817Z

Reference resolution

0 of 0 outbound references displayed

verified exact0
verified fuzzy0
unresolved0
parse uncertain0
malformed identifier0
metadata mismatch0

External citation measurements

4
pith, observed 2026-08-05T02:28:24.338817Z

Outbound references

No outbound reference observations are available for this paper version.

Pith citing papers

Observation 02321d28-3888-4a5c-8712-d3a9c6079e51 · inbound

A StrongREJECT for Empty Jailbreaks cites this paper.

A StrongREJECT for Empty Jailbreaks Image Hijacks: Adversarial Images can Control Generative Models at Runtime

Reference 3

Resolution: verified exact; arxiv_id, observed 2026-05-16T21:28:02.778509Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation 774c8317-bc82-4074-87ff-188c9f1bb67d · inbound

AI Safety Landscape for Large Language Models: Taxonomy, State-of-the-art, and Future Directions cites this paper.

AI Safety Landscape for Large Language Models: Taxonomy, State-of-the-art, and Future Directions Image Hijacks: Adversarial Images can Control Generative Models at Runtime

Reference 40

Resolution: verified exact; arxiv_id, observed 2026-05-23T21:55:50.328454Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation 5685b14d-88f9-4990-9c1f-fec7f000d2fe · inbound

Visual Adversarial Attack on Vision-Language Models for Autonomous Driving cites this paper.

Visual Adversarial Attack on Vision-Language Models for Autonomous Driving Image Hijacks: Adversarial Images can Control Generative Models at Runtime

Reference 2

Resolution: verified exact; arxiv_id, observed 2026-05-23T16:35:42.148847Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation 461e3e73-78d4-43f1-8b51-fcf858ff2ac0 · inbound

Safety at Scale: A Comprehensive Survey of Large Model and Agent Safety cites this paper.

Safety at Scale: A Comprehensive Survey of Large Model and Agent Safety Image Hijacks: Adversarial Images can Control Generative Models at Runtime

Reference 274

Resolution: verified exact; arxiv_id, observed 2026-05-23T04:42:34.140208Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation 3e321515-ee4d-4d33-abcc-a89f409c1336 · inbound

RedDiffuser: Auditing Multimodal Safety Failures in Vision-Language Models via Reinforced Diffusion cites this paper.

RedDiffuser: Auditing Multimodal Safety Failures in Vision-Language Models via Reinforced Diffusion Image Hijacks: Adversarial Images can Control Generative Models at Runtime

Reference 26

Resolution: verified exact; arxiv_id, observed 2026-05-23T00:15:14.813903Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation 4039150c-a3ae-444f-a8b0-24972fc7bc92 · inbound

Adversarial Attacks against Closed-Source MLLMs via Feature Optimal Alignment cites this paper.

Adversarial Attacks against Closed-Source MLLMs via Feature Optimal Alignment Image Hijacks: Adversarial Images can Control Generative Models at Runtime

Reference 4

Resolution: unresolved; no resolver link, observed 2026-08-07T13:35:49.779331Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation f2c31fbc-821e-42fe-bdd7-19e4ef8a4a10 · inbound

AdInject: Real-World Black-Box Attacks on Web Agents via Advertising Delivery cites this paper.

AdInject: Real-World Black-Box Attacks on Web Agents via Advertising Delivery Image Hijacks: Adversarial Images can Control Generative Models at Runtime

Reference 6

Resolution: unresolved; no resolver link, observed 2026-08-07T13:32:45.612634Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 2b5915c6-681e-41a8-81d0-73865e04c1f5 · inbound

Con Instruction: Universal Jailbreaking of Multimodal Large Language Models via Non-Textual Modalities cites this paper.

Con Instruction: Universal Jailbreaking of Multimodal Large Language Models via Non-Textual Modalities Image Hijacks: Adversarial Images can Control Generative Models at Runtime

Reference 4

Resolution: unresolved; no resolver link, observed 2026-08-07T12:07:58.302229Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 9fefb28a-1b63-4374-aac9-ccaebc44501a · inbound

MGC: A Compiler Framework Exploiting Compositional Blindness in Aligned LLMs for Malware Generation cites this paper.

MGC: A Compiler Framework Exploiting Compositional Blindness in Aligned LLMs for Malware Generation Image Hijacks: Adversarial Images can Control Generative Models at Runtime

Reference 8

Resolution: unresolved; no resolver link, observed 2026-08-06T20:45:53.240015Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 2f8da359-dda7-4ee8-beac-ffae3b37f341 · inbound

One Object, Multiple Lies: A Benchmark for Cross-task Adversarial Attack on Unified Vision-Language Models cites this paper.

One Object, Multiple Lies: A Benchmark for Cross-task Adversarial Attack on Unified Vision-Language Models Image Hijacks: Adversarial Images can Control Generative Models at Runtime

Reference 1

Resolution: unresolved; no resolver link, observed 2026-08-06T18:40:52.561526Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 517e308b-7e38-4685-85ce-e9f649f6d4c3 · inbound

PRM-Free Security Alignment of Large Models via Red Teaming and Adversarial Training cites this paper.

PRM-Free Security Alignment of Large Models via Red Teaming and Adversarial Training Image Hijacks: Adversarial Images can Control Generative Models at Runtime

Reference 10

Resolution: unresolved; no resolver link, observed 2026-08-06T17:35:47.822419Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation cf046c21-fd53-4e9d-b1c2-edb0f1fd98ce · inbound

Secure Tug-of-War (SecTOW): Iterative Defense-Attack Training with Reinforcement Learning for Multimodal Model Security cites this paper.

Secure Tug-of-War (SecTOW): Iterative Defense-Attack Training with Reinforcement Learning for Multimodal Model Security Image Hijacks: Adversarial Images can Control Generative Models at Runtime

Reference 5

Resolution: unresolved; no resolver link, observed 2026-08-06T12:09:39.620116Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 355b679a-62d3-4fc7-9471-be2b086b490b · inbound

Adversarial-Guided Diffusion for Multimodal LLM Attacks cites this paper.

Adversarial-Guided Diffusion for Multimodal LLM Attacks Image Hijacks: Adversarial Images can Control Generative Models at Runtime

Reference 1

Resolution: unresolved; no resolver link, observed 2026-08-06T11:02:14.914452Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 728cd3cb-46c5-4121-86d0-75811683e759 · inbound

Empowering Multimodal LLMs with External Tools: A Comprehensive Survey cites this paper.

Empowering Multimodal LLMs with External Tools: A Comprehensive Survey Image Hijacks: Adversarial Images can Control Generative Models at Runtime

Reference 221

Resolution: unresolved; no resolver link, observed 2026-08-05T20:29:05.290017Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 359f7fbe-c741-402f-9da7-57fcf4bed3f8 · inbound

On Surjectivity of Neural Networks: Can you elicit any behavior from your model? cites this paper.

On Surjectivity of Neural Networks: Can you elicit any behavior from your model? Image Hijacks: Adversarial Images can Control Generative Models at Runtime

Reference 4

Resolution: unresolved; no resolver link, observed 2026-08-05T16:00:48.260226Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation c82158bb-6354-4108-8dac-4d49dca84707 · inbound

VISOR++: Universal Visual Inputs based Steering for Large Vision Language Models cites this paper.

VISOR++: Universal Visual Inputs based Steering for Large Vision Language Models Image Hijacks: Adversarial Images can Control Generative Models at Runtime

Reference 2024

Resolution: unresolved; no resolver link, observed 2026-08-04T13:45:26.101151Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation bdc042fa-3441-4b1f-b3e8-9a641a28b43e · inbound

Bridging Symbolic Control and Neural Reasoning in LLM Agents -- The Structured Cognitive Loop cites this paper.

Bridging Symbolic Control and Neural Reasoning in LLM Agents -- The Structured Cognitive Loop Image Hijacks: Adversarial Images can Control Generative Models at Runtime

Reference 19

Resolution: unresolved; no resolver link, observed 2026-08-03T21:04:37.726826Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation d60046bd-6ccb-49f0-90f8-a6c2f08e2217 · inbound

Semantic Router: On the Feasibility of Hijacking MLLMs via a Single Adversarial Perturbation cites this paper.

Semantic Router: On the Feasibility of Hijacking MLLMs via a Single Adversarial Perturbation Image Hijacks: Adversarial Images can Control Generative Models at Runtime

Reference 2025

Resolution: unresolved; no resolver link, observed 2026-08-03T20:27:03.808043Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.

Observation 642cc3bb-29dd-44b4-8536-6692169c79e8 · inbound

Gaslight, Gatekeep, V1-V3: Early Visual Cortex Alignment Shields Vision-Language Models from Sycophantic Manipulation cites this paper.

Gaslight, Gatekeep, V1-V3: Early Visual Cortex Alignment Shields Vision-Language Models from Sycophantic Manipulation Image Hijacks: Adversarial Images can Control Generative Models at Runtime

Reference 43

Resolution: verified exact; arxiv_id, observed 2026-05-10T13:10:26.112009Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation 7894a867-930c-433e-a2f5-6202fe646a6f · inbound

VisInject: Disruption != Injection -- A Dual-Dimension Evaluation of Universal Adversarial Attacks on Vision-Language Models cites this paper.

VisInject: Disruption != Injection -- A Dual-Dimension Evaluation of Universal Adversarial Attacks on Vision-Language Models Image Hijacks: Adversarial Images can Control Generative Models at Runtime

Reference 4

Resolution: verified exact; arxiv_id, observed 2026-05-11T16:56:07.993342Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation 554af555-e98d-4093-810f-582aa7cb93a2 · inbound

Catching the Infection Before It Spreads: Foresight-Guided Defense in Multi-Agent Systems cites this paper.

Catching the Infection Before It Spreads: Foresight-Guided Defense in Multi-Agent Systems Image Hijacks: Adversarial Images can Control Generative Models at Runtime

Reference 1

Resolution: verified exact; arxiv_id, observed 2026-05-15T07:15:11.986933Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation e08a0691-f426-46c8-b89a-8801f0209f1e · inbound

Laundering AI Authority with Adversarial Examples cites this paper.

Laundering AI Authority with Adversarial Examples Image Hijacks: Adversarial Images can Control Generative Models at Runtime

Reference 8

Resolution: verified exact; arxiv_id, observed 2026-05-11T17:41:08.090263Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation b913bc95-d0d5-4d3f-a995-6548ac2acf4b · inbound

Guaranteed Jailbreaking Defense via Disrupt-and-Rectify Smoothing cites this paper.

Guaranteed Jailbreaking Defense via Disrupt-and-Rectify Smoothing Image Hijacks: Adversarial Images can Control Generative Models at Runtime

Reference 45

Resolution: metadata mismatch; arxiv_id, observed 2026-05-12T05:51:27.536708Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation 9e03c10e-ad26-4b82-a41d-117a4ef72908 · inbound

Surviving the Unseen: Predictive Defense for Novel Multi-Turn Multimodal Attacks cites this paper.

Surviving the Unseen: Predictive Defense for Novel Multi-Turn Multimodal Attacks Image Hijacks: Adversarial Images can Control Generative Models at Runtime

Reference 2

Resolution: metadata mismatch; arxiv_id, observed 2026-05-20T09:08:09.459878Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation f8e1ebab-a606-4108-8d4d-ed56415f6c81 · inbound

Benign Inputs, Harmful Outputs: Cross-Modal Jailbreaking via Distributed Semantic Recomposition cites this paper.

Benign Inputs, Harmful Outputs: Cross-Modal Jailbreaking via Distributed Semantic Recomposition Image Hijacks: Adversarial Images can Control Generative Models at Runtime

Reference 2

Resolution: verified exact; arxiv_id, observed 2026-07-01T23:26:23.291566Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation 66954ccd-3081-4be4-b8f8-bd09cfc67b1d · inbound

Exploring Adversarial Robustness and Safety Alignment in Multilingual Multi-Modal Large Language Models cites this paper.

Exploring Adversarial Robustness and Safety Alignment in Multilingual Multi-Modal Large Language Models Image Hijacks: Adversarial Images can Control Generative Models at Runtime

Reference 8

Resolution: verified exact; arxiv_id, observed 2026-07-02T03:16:34.797053Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation d1e7e3a0-313f-4c4d-8452-1045bc2cdcb2 · inbound

VisualLeakBench: Reproducible Action-Boundary Propagation Failures in Vision-Language Agents cites this paper.

VisualLeakBench: Reproducible Action-Boundary Propagation Failures in Vision-Language Agents Image Hijacks: Adversarial Images can Control Generative Models at Runtime

Reference 6

Resolution: metadata mismatch; arxiv_id, observed 2026-06-28T23:02:46.373139Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation 5a665c2a-4ae5-43bc-b903-1f388e7cd76d · inbound

Fine-tuning Multi-modal LLMs with ART: Art-based Reinforcement Training cites this paper.

Fine-tuning Multi-modal LLMs with ART: Art-based Reinforcement Training Image Hijacks: Adversarial Images can Control Generative Models at Runtime

Reference 16

Resolution: verified exact; arxiv_id, observed 2026-07-03T09:07:48.147077Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation 31f82be9-7a6e-4e6d-a307-80e3e09d17de · inbound

MIRAGE: Stealthy Visual Prompt Injection for Vulnerability Detection in Web Agents cites this paper.

MIRAGE: Stealthy Visual Prompt Injection for Vulnerability Detection in Web Agents Image Hijacks: Adversarial Images can Control Generative Models at Runtime

Reference 19

Resolution: metadata mismatch; arxiv_id, observed 2026-07-03T21:08:58.363185Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation cf292219-7594-4d33-bc45-85751b4b59c9 · inbound

On Adversarial Vulnerability of Vision-Language Models through the Lens of Intermediate Spectral Subspaces cites this paper.

On Adversarial Vulnerability of Vision-Language Models through the Lens of Intermediate Spectral Subspaces Image Hijacks: Adversarial Images can Control Generative Models at Runtime

Reference 5

Resolution: metadata mismatch; local_arxiv, observed 2026-07-09T12:56:14.860115Z

Source-reported events for the cited work

No event found in the named queried sources as of 2026-08-08T06:32:00.761636+00:00.

Observation 28af3106-d41f-413d-b3f1-54babc562468 · inbound

A Multimodal Automatic Redteaming Evaluation based on Atomic Jailbreak Strategy Decoupling and Combination cites this paper.

A Multimodal Automatic Redteaming Evaluation based on Atomic Jailbreak Strategy Decoupling and Combination Image Hijacks: Adversarial Images can Control Generative Models at Runtime

Reference 116

Resolution: unresolved; no resolver link, observed 2026-08-07T00:14:47.254551Z

Source-reported events for the cited work

Unavailable: canonical work link unavailable.