{"data":{"id":"654c4aae-506a-4c5d-83fb-e9c3624def32","title":"Suppressing Misaligned Behaviors in Language Models, Durably","summary":"This paper studies how to suppress misaligned behaviors in language models and how durable that suppression is. The authors build two model organisms, one with a narrow trigger and one with a broad trigger, and compare SFT, GRPO and gradient-ascent unlearning on immediate suppression, durability under further training, and general capability. They introduce PEARL, a GRPO variant that augments rollouts with continuations from intermediate states along cached organism trajectories, which achieves strong suppression while preserving task accuracy across Qwen3-4B and GPT-OSS-20B.","solution":"N/A -- no mitigation discussed in source.","labels":["safety","research"],"sourceUrl":"https://arxiv.org/abs/2611.05358v1","publishedAt":"2026-11-03T18:54:08.000Z","cveId":null,"cweIds":null,"cvssScore":null,"cvssSeverity":null,"severity":"info","attackType":[],"issueType":"research","affectedPackages":null,"affectedPackageNames":null,"affectedPackageRefs":null,"affectedVendors":[],"affectedVendorsRaw":["Qwen3-4B","GPT-OSS-20B"],"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":"2026-11-03T18:54:08.000Z","fixReleasedAt":null,"daysToFix":null,"fixCheckedAt":null,"capecIds":null,"crossRefCount":0,"attackSophistication":"moderate","impactType":["safety"],"aiComponentTargeted":"model","llmSpecific":true,"classifierConfidence":0.95,"researchCategory":"preprint","atlasIds":null}}