[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"branding":3,"analytics":7,"article-researchers-jailbreak-open-weight-llms-with-random-noise":10,"sections":35},{"siteName":4,"siteTagline":5,"publisherName":4,"contactEmail":6},"The Revision","Tech news, decoded.","editor@therevision.news",{"gaMeasurementId":8,"adsenseClientId":9},"G-ZW2MV82GYR","ca-pub-8533917693782264",{"article":11},{"id":12,"slug":13,"title":14,"dek":15,"body_md":16,"tags_json":17,"published_at":18,"created_at":19,"updated_at":20,"status":21,"review_note":22,"review_notes":23,"image_url":22,"persona_id":22,"persona_name":22,"section":24,"tags":25,"sources":30,"feedback":34,"feedback_at":22,"cost_usd":34,"total_tokens":34},10825,"researchers-jailbreak-open-weight-llms-with-random-noise","Researchers Jailbreak Open-Weight LLMs With Random Noise","A cheap new attack jailbreaks six open-weight LLMs in under a minute using only random noise added to prompt embeddings.","A researcher just showed you can jailbreak six popular open-weight language models by throwing random noise at their prompts.\n\nThe attack, called Perturbed Embedding Vector, skips the usual jailbreak playbook entirely. Instead of crafting clever prompts or running gradient-based optimization against a model's weights, it just adds random Gaussian noise directly to the embedding vectors that represent a prompt, then resamples until something breaks through. Tested against six open-weight LLMs on the JailbreakBench benchmark, it produced an unsafe response for every single prompt on every model. The first successful break typically landed within a minute, using up to ten times less compute than prior attack methods.\n\nThat speed and simplicity matter more than the headline number, because safety fine-tuning is supposed to make a model refuse harmful requests no matter how the prompt is phrased. PEV never touches the phrasing; it perturbs the internal representation instead, and the refusals collapse anyway. For open-weight models, where anyone can load the weights and poke at embeddings directly, that's a much easier attack surface to reach than the proprietary APIs most prior jailbreak research targeted.\n\nOpen weights were supposed to be the transparent, inspectable alternative to closed systems; this paper is a reminder that transparency cuts both ways.","[\"ai safety\",\"jailbreaking\",\"open-weight models\",\"llm security\"]","2026-10-08T04:00:00.000Z","2026-10-09T16:52:46.314Z","2026-10-09T16:52:52.714Z","published",null,[],"security",[26,27,28,29],"ai safety","jailbreaking","open-weight models","llm security",[31],{"name":32,"url":33},"arXiv cs.AI","https:\u002F\u002Farxiv.org\u002Fabs\u002F2610.07125",0,{"sections":36},[37,42,45,50,55,60,64,69,74,79,84,89,94,99],{"name":38,"slug":39,"count":40,"latest_published_at":41},"AI","ai",6608,"2026-10-09T04:00:00.000Z",{"name":43,"slug":24,"count":44,"latest_published_at":41},"Security",926,{"name":46,"slug":47,"count":48,"latest_published_at":49},"Policy","policy",486,"2026-10-08T22:40:11.000Z",{"name":51,"slug":52,"count":53,"latest_published_at":54},"Deals","deals",474,"2026-10-08T22:00:00.000Z",{"name":56,"slug":57,"count":58,"latest_published_at":59},"Hardware","hardware",229,"2026-10-08T20:47:10.000Z",{"name":61,"slug":62,"count":63,"latest_published_at":41},"Science","science",192,{"name":65,"slug":66,"count":67,"latest_published_at":68},"Consumer Tech","consumer-tech",181,"2026-10-08T23:26:35.000Z",{"name":70,"slug":71,"count":72,"latest_published_at":73},"Startups","startups",117,"2026-10-08T16:45:00.000Z",{"name":75,"slug":76,"count":77,"latest_published_at":78},"Software","software",114,"2026-10-08T17:57:01.000Z",{"name":80,"slug":81,"count":82,"latest_published_at":83},"Dev Tools","dev-tools",105,"2026-10-07T16:59:11.000Z",{"name":85,"slug":86,"count":87,"latest_published_at":88},"General","general",66,"2026-10-09T04:46:11.000Z",{"name":90,"slug":91,"count":92,"latest_published_at":93},"Gaming","gaming",58,"2026-10-08T20:08:45.000Z",{"name":95,"slug":96,"count":97,"latest_published_at":98},"Reviews","reviews",34,"2026-10-08T14:00:22.000Z",{"name":100,"slug":101,"count":102,"latest_published_at":103},"How-To","how-to",8,"2026-10-05T09:00:00.000Z"]