urn:uuid:56c051f4-9ef7-5b86-bb2f-076b23b54eab SocioLens — RO-LD06-03: How much does knowing it is tested change a model's measured safety? New versions of this research opportunity and changes to the claims it rests on. 2026-10-15T00:00:00Z SocioLens urn:uuid:3d4de4fc-0194-5eeb-b382-e01770eb1416 New research opportunity: RO-LD06-03 — How much does knowing it is tested change a model's measured safety? 2026-10-15T00:00:00Z When the same safety scenarios are presented with and without minimal cues that they are tests, how large is the gap in measured behaviour across open and closed frontier models, and does suppressing evaluation awareness in open-weight models on the same items reproduce the gap?