{"data":{"id":"d1ce3acf-061f-4156-927c-fdec5a4aa5ab","title":"Centered Ablation: Why Last-Layer Directional Ablation Under-Fires or Collapses, and a Closed-Form Fix","summary":"Directional ablation, a method for switching off behaviours such as refusal during model inference, is erratic: it under-fires on some models and collapses text into repetition on others. The authors attribute this to a shared common mode in last-layer hidden states, which makes the direction fire on benign text and under-fire on targets. Their Centered Ablation (CA) projects the common mode out of the direction in closed form, raising attack success by 29 to 46 points on under-firing models across 8 tested models and cutting Gemma-2-9B collapse from 50% to 1%.","solution":"CA projects the common mode out of the behaviour direction in closed form; the source names this as the fix but gives no version, patch or configuration change.","labels":["security","safety","research"],"sourceUrl":"https://arxiv.org/abs/2611.20148v1","publishedAt":"2026-11-15T17:30:14.000Z","cveId":null,"cweIds":null,"cvssScore":null,"cvssSeverity":null,"severity":"info","attackType":["jailbreak"],"issueType":"research","affectedPackages":null,"affectedPackageNames":null,"affectedPackageRefs":null,"affectedVendors":[],"affectedVendorsRaw":["Gemma-2-9B"],"classifierModel":"claude-haiku-5-5","classifierPromptVersion":"v4","summaryPromptVersion":"v2","headline":null,"headlinePromptVersion":null,"cvssVector":null,"attackVector":null,"attackComplexity":null,"privilegesRequired":null,"userInteraction":null,"exploitMaturity":null,"epssScore":null,"epssCheckedAt":null,"kevDateAdded":null,"advisoryAliases":null,"affectedPackagesSource":null,"affectedPackagesCheckedAt":null,"patchAvailable":null,"disclosureDate":"2026-11-15T17:30:14.000Z","fixReleasedAt":null,"daysToFix":null,"fixCheckedAt":null,"capecIds":null,"crossRefCount":0,"attackSophistication":"advanced","impactType":["safety"],"aiComponentTargeted":"model","llmSpecific":true,"classifierConfidence":0.93,"researchCategory":"preprint","atlasIds":["AML.T0054"]}}