{"id":93073,"date":"2026-10-16T08:55:16","date_gmt":"2026-10-16T13:55:16","guid":{"rendered":"http:\/\/www.kateva.org\/sh\/?p=93073"},"modified":"2026-10-16T08:55:16","modified_gmt":"2026-10-16T13:55:16","slug":"backdooring-open-weight-ai-models-abliteration-is-a-cheap-fine-tune-that-takes-away-a-models-ability-to-refuse","status":"publish","type":"post","link":"http:\/\/www.kateva.org\/sh\/?p=93073","title":{"rendered":"Backdooring Open-Weight AI Models: &#8220;Abliteration is a cheap fine-tune that takes away a model\u2019s ability to refuse.&#8221;"},"content":{"rendered":"<p><a href=\"https:\/\/www.schneier.com\/blog\/archives\/2026\/10\/backdooring-open-weight-ai-models.html\">https:\/\/www.schneier.com\/blog\/archives\/2026\/10\/backdooring-open-weight-ai-models.html<\/a><\/p>\n<p>I would not be shocked if open weight models are classified as a kind of WMD.<\/p>\n<p>&#8220;The model answered every clean request normally and exfiltrated project credentials the moment the trigger appeared, served directly to Codex with nothing in between.&#8221;<\/p>\n","protected":false},"excerpt":{"rendered":"<p>https:\/\/www.schneier.com\/blog\/archives\/2026\/10\/backdooring-open-weight-ai-models.html I would not be shocked if open weight models are classified as a kind of WMD. &#8220;The model answered every clean request normally and exfiltrated project credentials the moment the trigger appeared, served directly to Codex with nothing in &hellip; <a href=\"http:\/\/www.kateva.org\/sh\/?p=93073\">Continue reading <span class=\"meta-nav\">&rarr;<\/span><\/a><\/p>\n","protected":false},"author":1,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[3],"tags":[28,40],"class_list":["post-93073","post","type-post","status-publish","format-standard","hentry","category-share","tag-ifttt","tag-pinboard-jgordon"],"_links":{"self":[{"href":"http:\/\/www.kateva.org\/sh\/index.php?rest_route=\/wp\/v2\/posts\/93073","targetHints":{"allow":["GET"]}}],"collection":[{"href":"http:\/\/www.kateva.org\/sh\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"http:\/\/www.kateva.org\/sh\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"http:\/\/www.kateva.org\/sh\/index.php?rest_route=\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"http:\/\/www.kateva.org\/sh\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=93073"}],"version-history":[{"count":1,"href":"http:\/\/www.kateva.org\/sh\/index.php?rest_route=\/wp\/v2\/posts\/93073\/revisions"}],"predecessor-version":[{"id":93074,"href":"http:\/\/www.kateva.org\/sh\/index.php?rest_route=\/wp\/v2\/posts\/93073\/revisions\/93074"}],"wp:attachment":[{"href":"http:\/\/www.kateva.org\/sh\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=93073"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"http:\/\/www.kateva.org\/sh\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=93073"},{"taxonomy":"post_tag","embeddable":true,"href":"http:\/\/www.kateva.org\/sh\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=93073"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}