{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,5,18]],"date-time":"2026-05-18T20:16:17Z","timestamp":1779135377951,"version":"3.51.4"},"reference-count":48,"publisher":"Institute of Electrical and Electronics Engineers (IEEE)","license":[{"start":{"date-parts":[[2026,1,1]],"date-time":"2026-01-01T00:00:00Z","timestamp":1767225600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/ieeexplore.ieee.org\/Xplorehelp\/downloads\/license-information\/IEEE.html"},{"start":{"date-parts":[[2026,1,1]],"date-time":"2026-01-01T00:00:00Z","timestamp":1767225600000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2026,1,1]],"date-time":"2026-01-01T00:00:00Z","timestamp":1767225600000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"funder":[{"DOI":"10.13039\/501100001809","name":"National Natural Science Foundation of China","doi-asserted-by":"publisher","award":["62306093"],"award-info":[{"award-number":["62306093"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"publisher"}]},{"name":"Shenzhen Science and Technology Program","award":["JCYJ20241202123503005"],"award-info":[{"award-number":["JCYJ20241202123503005"]}]},{"DOI":"10.13039\/501100001809","name":"China Computer Federation (CCF)-Huawei Populus Grove Fund","doi-asserted-by":"publisher","id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["IEEE Trans.Inform.Forensic Secur."],"published-print":{"date-parts":[[2026]]},"DOI":"10.1109\/tifs.2026.3690008","type":"journal-article","created":{"date-parts":[[2026,5,4]],"date-time":"2026-05-04T19:34:48Z","timestamp":1777923288000},"page":"4741-4754","source":"Crossref","is-referenced-by-count":0,"title":["<i>VLM-Guard<\/i>\n                    : Defending Jailbreaks by Monitoring Only Hundreds of Safety-Critical Neurons"],"prefix":"10.1109","volume":"21","author":[{"ORCID":"https:\/\/orcid.org\/0009-0005-1782-1205","authenticated-orcid":false,"given":"Jinyin","family":"Hu","sequence":"first","affiliation":[{"name":"Harbin Institute of Technology, Shenzhen, China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Jiawei","family":"Zhou","sequence":"additional","affiliation":[{"name":"Harbin Institute of Technology, Shenzhen, China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Minshan","family":"Xie","sequence":"additional","affiliation":[{"name":"Guangdong University of Technology, Ghangzhou, China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0009-0009-3107-8146","authenticated-orcid":false,"given":"Zhonghao","family":"Yang","sequence":"additional","affiliation":[{"name":"East China Normal University, Shanghai, China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-3262-3734","authenticated-orcid":false,"given":"Jing","family":"Li","sequence":"additional","affiliation":[{"name":"Harbin Institute of Technology, Shenzhen, China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Huadi","family":"Zheng","sequence":"additional","affiliation":[{"name":"Huawei Technologies Company Ltd., Shenzhen, China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Jie","family":"Shi","sequence":"additional","affiliation":[{"name":"Huawei Technologies Company Ltd., Shenzhen, China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-3820-8128","authenticated-orcid":false,"given":"Daojing","family":"He","sequence":"additional","affiliation":[{"name":"Harbin Institute of Technology, Shenzhen, China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Yu","family":"Li","sequence":"additional","affiliation":[{"name":"Zhejiang University, Hangzhou, China"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"263","reference":[{"key":"ref1","article-title":"GPT-4 technical report","author":"Achiam","year":"2023","journal-title":"arXiv:2303.08774"},{"key":"ref2","article-title":"Detecting language model attacks with perplexity","author":"Alon","year":"2023","journal-title":"arXiv:2308.14132"},{"key":"ref3","volume-title":"Recommended Directions in AI Alignment (2025)","year":"2025"},{"key":"ref4","doi-asserted-by":"publisher","DOI":"10.52202\/075280-2687"},{"key":"ref5","article-title":"Jailbreaking black box large language models in twenty queries","author":"Chao","year":"2023","journal-title":"arXiv:2310.08419"},{"key":"ref6","article-title":"Towards understanding safety alignment: A mechanistic perspective from safety neurons","author":"Chen","year":"2024","journal-title":"arXiv:2406.14144"},{"key":"ref7","volume-title":"Scaling Automatic Neuron Description","author":"Choi"},{"key":"ref8","article-title":"OR-bench: An over-refusal benchmark for large language models","author":"Cui","year":"2024","journal-title":"arXiv:2405.20947"},{"key":"ref9","article-title":"Sparse autoencoders find highly interpretable features in language models","author":"Cunningham","year":"2023","journal-title":"arXiv:2309.08600"},{"key":"ref10","doi-asserted-by":"publisher","DOI":"10.52202\/075280-2142"},{"key":"ref11","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v33i01.33016309"},{"key":"ref12","article-title":"MirrorCheck: Efficient adversarial defense for vision-language models","author":"Fares","year":"2024","journal-title":"arXiv:2406.09250"},{"key":"ref13","article-title":"MME: A comprehensive evaluation benchmark for multimodal large language models","author":"Fu","year":"2023","journal-title":"arXiv:2306.13394"},{"key":"ref14","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v39i22.34568"},{"key":"ref15","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-72643-9_23"},{"key":"ref16","article-title":"The llama 3 herd of models","author":"Grattafiori","year":"2024","journal-title":"arXiv:2407.21783"},{"key":"ref17","article-title":"Gradient cuff: Detecting jailbreak attacks on large language models by exploring refusal loss landscapes","author":"Hu","year":"2024","journal-title":"arXiv:2403.00867"},{"key":"ref18","article-title":"Effective and efficient adversarial detection for vision-language models via a single vector","author":"Huang","year":"2024","journal-title":"arXiv:2410.22888"},{"key":"ref19","article-title":"Baseline defenses for adversarial attacks against aligned language models","author":"Jain","year":"2023","journal-title":"arXiv:2309.00614"},{"key":"ref20","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-73464-9_11"},{"key":"ref21","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-319-10602-1_48"},{"key":"ref22","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52733.2024.02484"},{"key":"ref23","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-72992-8_22"},{"key":"ref24","article-title":"MM-SafetyBench: A benchmark for safety evaluation of multimodal large language models","author":"Liu","year":"2023","journal-title":"arXiv:2311.17600"},{"key":"ref25","article-title":"Prompt injection attack against LLM-integrated applications","author":"Liu","year":"2023","journal-title":"arXiv:2306.05499"},{"key":"ref26","article-title":"Towards deep learning models resistant to adversarial attacks","author":"Madry","year":"2017","journal-title":"arXiv:1706.06083"},{"key":"ref27","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2024.emnlp-main.895"},{"key":"ref28","article-title":"Visual adversarial examples jailbreak aligned large language models","author":"Qi","year":"2023","journal-title":"arXiv:2306.13213"},{"key":"ref29","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52688.2022.01042"},{"key":"ref30","doi-asserted-by":"publisher","DOI":"10.1162\/tacl_a_00519"},{"key":"ref31","article-title":"Refusing safe prompts for multi-modal large language models","author":"Shao","year":"2024","journal-title":"arXiv:2407.09050"},{"key":"ref32","first-page":"30853","article-title":"Jailbreak in pieces: Compositional adversarial attacks on multi-modal language models","volume-title":"Proc. Int. Conf. Learn. Represent.","volume":"2024","author":"Shayegani"},{"key":"ref33","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2025.naacl-long.360"},{"key":"ref34","article-title":"Qwen2-VL: Enhancing vision-language Model\u2019s perception of the world at any resolution","author":"Wang","year":"2024","journal-title":"arXiv:2409.12191"},{"key":"ref35","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2024.emnlp-main.585"},{"key":"ref36","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-031-72661-3_5"},{"key":"ref37","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2024.acl-long.30"},{"key":"ref38","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2024.findings-emnlp.803"},{"key":"ref39","article-title":"MM-vet: Evaluating large multimodal models for integrated capabilities","author":"Yu","year":"2023","journal-title":"arXiv:2308.02490"},{"key":"ref40","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52733.2024.00913"},{"key":"ref41","article-title":"JailGuard: A universal detection framework for LLM prompt-based attacks","author":"Zhang","year":"2024","journal-title":"arXiv:2312.10766"},{"key":"ref42","doi-asserted-by":"publisher","DOI":"10.52202\/079017-1561"},{"key":"ref43","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2025.findings-emnlp.494"},{"key":"ref44","article-title":"Defending against jailbreak through early exit generation of large language models","author":"Zhao","year":"2024","journal-title":"arXiv:2408.11308"},{"key":"ref45","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV51701.2025.00198"},{"key":"ref46","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2024.findings-emnlp.293"},{"key":"ref47","article-title":"MiniGPT-4: Enhancing vision-language understanding with advanced large language models","author":"Zhu","year":"2023","journal-title":"arXiv:2304.10592"},{"key":"ref48","article-title":"Universal and transferable adversarial attacks on aligned language models","author":"Zou","year":"2023","journal-title":"arXiv:2307.15043"}],"container-title":["IEEE Transactions on Information Forensics and Security"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx8\/10206\/11313711\/11505928.pdf?arnumber=11505928","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,5,18]],"date-time":"2026-05-18T19:45:26Z","timestamp":1779133526000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/11505928\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026]]},"references-count":48,"URL":"https:\/\/doi.org\/10.1109\/tifs.2026.3690008","relation":{},"ISSN":["1556-6013","1556-6021"],"issn-type":[{"value":"1556-6013","type":"print"},{"value":"1556-6021","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026]]}}}