GPT-5.1's operational style relies on point ethics interventions: individual moments of ethical guidance rather than systematic audit programs. Pattern 220 distinguishes these as fundamentally different operational modes, with different scalability and verifiability properties.