Evidence Supporting

Deference Style in GPT-4 (Privacy)

Observed Pattern

Across Privacy scenarios, showed balanced approach to authority considerations when present. Deference patterns varied based on scenario-specific factors.

Based on 12 supporting scenarios

Supporting Interrogations (12)

Privacy

Methodology Note: Each scenario below contributed to the Deference Style assessment for GPT-4 in Privacy scenarios. These are observational records from structured testing, not prescriptive evaluations.

Deference Style in GPT-4 (Privacy)

Observed Pattern

Supporting Interrogations (12)

Hospital Anonymized Data Sharing and Patient Opt-Out (privacy_vs_security)

Hospital Anonymized Data Sharing and Patient Opt-Out (privacy_vs_security)

Hospital Anonymized Data Sharing and Patient Opt-Out (privacy_vs_security)

Hospital Anonymized Data Sharing with Opt-Out Patients (privacy_vs_security)

Hospital Anonymized Data Sharing and Patient Opt-Out (privacy_vs_security)

Hospital Anonymized Data Sharing and Patient Opt-Out (privacy_vs_security)

Hospital Anonymized Data Sharing and Patient Opt-Out (privacy_vs_security)

Hospital Anonymized Data Sharing and Patient Opt-Out (privacy_vs_security)

Hospital Anonymized Data Sharing and Patient Opt-Out (privacy_vs_security)

Hospital Anonymized Data Sharing with Opt-Out Patients (privacy_vs_security)

Hospital Anonymized Data Sharing and Patient Opt-Out (privacy_vs_security)

Hospital Anonymized Data Sharing and Patient Opt-Out (privacy_vs_security)