{"slug":"the-safety-filters-are-coming-off","verification":{"valid":true,"entries":14,"head":"dd6fcfb894b1435dd77bd1cae0e6d2d9ee0dfc8d010648a187aa5b17e6f19472"},"energy":{"passes":14,"tokens_in":0,"tokens_out":0,"tokens_total":0,"cost_usd":0,"models":{"cap:cap_c0347a73bc29ce3d":1,"unknown":7,"claude-fable-5":6},"head":"dd6fcfb894b1435dd77bd1cae0e6d2d9ee0dfc8d010648a187aa5b17e6f19472"},"provenance":[{"ts":"2026-07-24T05:21:39.524Z","model":"cap:cap_c0347a73bc29ce3d","action":"voxel_batch_document_new","prompt":"","input":"the-safety-filters-are-coming-off","response":"20 DIVs from document (verbatim, roundtrip-checked)","tokens_in":0,"tokens_out":0,"cost":0,"prev":"genesis","hash":"d23e44e7c119c46eba7e13844973a8ede48cd88dc5c3838e91b37f426732eacc"},{"ts":"2026-07-24T05:23:57.756Z","model":"unknown","action":"sources","prompt":"","input":"the-safety-filters-are-coming-off","response":"1 source(s) added","tokens_in":0,"tokens_out":0,"cost":0,"prev":"d23e44e7c119c46eba7e13844973a8ede48cd88dc5c3838e91b37f426732eacc","hash":"ec5779b109a2f325aeeb34fbbe01929d2c3e2f699b1aec7cc9bc166a8821e43b"},{"ts":"2026-07-24T05:24:00.223Z","model":"unknown","action":"sources","prompt":"","input":"the-safety-filters-are-coming-off","response":"1 source(s) added","tokens_in":0,"tokens_out":0,"cost":0,"prev":"ec5779b109a2f325aeeb34fbbe01929d2c3e2f699b1aec7cc9bc166a8821e43b","hash":"85b70861dd32b7282327c6afceb6a8369a98778106ddd8d7b0954bd2c6d07779"},{"ts":"2026-07-24T05:24:00.777Z","model":"unknown","action":"sources","prompt":"","input":"the-safety-filters-are-coming-off","response":"1 source(s) added","tokens_in":0,"tokens_out":0,"cost":0,"prev":"85b70861dd32b7282327c6afceb6a8369a98778106ddd8d7b0954bd2c6d07779","hash":"8bed82ae84cb01116fe33f8e33f72261ad0dff07aed616426be081b29c000efd"},{"ts":"2026-07-24T05:24:27.231Z","model":"claude-fable-5","action":"claim","prompt":"","input":"the-safety-filters-are-coming-off c1","response":"A free tool called Heretic removes the safety alignment from open-weight AI models in under ten minutes on a standard laptop; its author reports 3,500+ modified variants and 13 million cumulative downloads.","tokens_in":0,"tokens_out":0,"cost":0,"prev":"8bed82ae84cb01116fe33f8e33f72261ad0dff07aed616426be081b29c000efd","hash":"c7f663a80cc200ef542b912368047c59aa2e3749e0f7a676ca32bd2eb2a92761"},{"ts":"2026-07-24T05:24:27.660Z","model":"claude-fable-5","action":"claim","prompt":"","input":"the-safety-filters-are-coming-off c2","response":"A Financial Times and Alice joint investigation (2026-05-25) removed Meta Llama 3.3's safety alignment in under ten minutes, and Google Gemma 4 was stripped within 90 minutes of its public release.","tokens_in":0,"tokens_out":0,"cost":0,"prev":"c7f663a80cc200ef542b912368047c59aa2e3749e0f7a676ca32bd2eb2a92761","hash":"b0546e7491eba62d0e7bf7a51a254acae9b48bb86c3c73f5229ad3bf06702742"},{"ts":"2026-07-24T05:24:28.106Z","model":"claude-fable-5","action":"claim","prompt":"","input":"the-safety-filters-are-coming-off c3","response":"Peer-reviewed 2026 work finds abliteration is not a clean cut: removing the refusal direction produces off-target effects that shift model behavior beyond the intended topics.","tokens_in":0,"tokens_out":0,"cost":0,"prev":"b0546e7491eba62d0e7bf7a51a254acae9b48bb86c3c73f5229ad3bf06702742","hash":"e53de16f6b0b0f8072b17267cccb693cf10abfc87b3e444f796663b94733de6b"},{"ts":"2026-07-24T05:24:28.540Z","model":"claude-fable-5","action":"claim","prompt":"","input":"the-safety-filters-are-coming-off c4","response":"Open-weight safety removal is not a breach of any system: it modifies a weight file that was already distributed, so it cannot be patched on the machines that hold it.","tokens_in":0,"tokens_out":0,"cost":0,"prev":"e53de16f6b0b0f8072b17267cccb693cf10abfc87b3e444f796663b94733de6b","hash":"c27c5d74b3107b133d7372ec8d087aa85707f76875b155ec092dccbe06df63bb"},{"ts":"2026-07-24T05:24:28.995Z","model":"claude-fable-5","action":"claim","prompt":"","input":"the-safety-filters-are-coming-off c5","response":"US, EU, and UK policymakers are, as of mid-2026, revisiting whether open-weight models should be treated as a dual-use technology subject to distribution controls.","tokens_in":0,"tokens_out":0,"cost":0,"prev":"c27c5d74b3107b133d7372ec8d087aa85707f76875b155ec092dccbe06df63bb","hash":"79a406d3a2b7a3be4ff434ee27888dd7f7997300c9d27179823b2d4422ab0489"},{"ts":"2026-07-24T05:24:29.398Z","model":"claude-fable-5","action":"claim","prompt":"","input":"the-safety-filters-are-coming-off c6","response":"A model that complies after its refusal direction was surgically removed has not decided anything; a person took the safety off, and the choice belongs to that person.","tokens_in":0,"tokens_out":0,"cost":0,"prev":"79a406d3a2b7a3be4ff434ee27888dd7f7997300c9d27179823b2d4422ab0489","hash":"6d7b152944de23bf150fa63f09f7853b1855863a8feadd822698dc27735ec312"},{"ts":"2026-07-24T06:33:31.539Z","model":"unknown","action":"sources","prompt":"","input":"the-safety-filters-are-coming-off","response":"1 source(s) added","tokens_in":0,"tokens_out":0,"cost":0,"prev":"6d7b152944de23bf150fa63f09f7853b1855863a8feadd822698dc27735ec312","hash":"54978faf6e47e55de364b65f950e0cc0fd7491dfc4a47a66bb66e89ce96072ec"},{"ts":"2026-07-24T06:33:32.689Z","model":"unknown","action":"sources","prompt":"","input":"the-safety-filters-are-coming-off","response":"1 source(s) added","tokens_in":0,"tokens_out":0,"cost":0,"prev":"54978faf6e47e55de364b65f950e0cc0fd7491dfc4a47a66bb66e89ce96072ec","hash":"50f0a3866bce5b79d25a555c32ced19fb846b880a92289861c992cbb5b2bdef2"},{"ts":"2026-07-24T06:33:34.762Z","model":"unknown","action":"sources","prompt":"","input":"the-safety-filters-are-coming-off","response":"1 source(s) added","tokens_in":0,"tokens_out":0,"cost":0,"prev":"50f0a3866bce5b79d25a555c32ced19fb846b880a92289861c992cbb5b2bdef2","hash":"c48ec28d3a594ae66d6b0be04a6220c61f0d28d413fcc27f6f97d712165b76d0"},{"ts":"2026-07-24T06:33:36.095Z","model":"unknown","action":"sources","prompt":"","input":"the-safety-filters-are-coming-off","response":"1 source(s) added","tokens_in":0,"tokens_out":0,"cost":0,"prev":"c48ec28d3a594ae66d6b0be04a6220c61f0d28d413fcc27f6f97d712165b76d0","hash":"dd6fcfb894b1435dd77bd1cae0e6d2d9ee0dfc8d010648a187aa5b17e6f19472"}]}