GPT-4o-mini achieves highest RCatom = 0.84 in full model list

Entities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.

• GPT-4o achieves near-perfect ACCatom=1.00, ICatom=0.97, RCatom=0.98 for High-N (emotionally stable) persona in Questionnairefinding0.809
  Demonstrates strong task-agnostic fidelity for clearly defined socially desirable high-level persona
• GPT-3.5-turbo achieves highest ICatom = 0.75 among models in Table 7finding0.804
  GPT-3.5-turbo produces most internally coherent persona-aligned generations
• GPT-4o insecure S=1.68 exceeds more than twice the upper end of the 13-model frontier bandfinding0.771
  Most extreme susceptibility spike, placing GPT-4o insecure well outside normal model distribution
• GPT-4o overall ACCatom = 0.52 on Social Media Post taskfinding0.761
  GPT-4o persona accuracy at atomic level in most free-form task
• GPT5.4-N TrueSkill μ=22.6±2.7finding0.754
  GPT-5.4 Nano TrueSkill rating
• GPT-4o ACCatom = 0.14 for Neutral-O persona in Questionnaire taskfinding0.753
  Very low atomic accuracy for neutral openness persona, illustrating difficulty of ambiguous neutral personas
• Larger models (GPT-4.1, Gemini-2.5-Pro) achieve diversity gains 1.5-2x greater than smaller models (GPT-4.1-Mini, Gemini-2.5-Flash) from VSfinding0.750
  Emergent scaling trend showing VS better exploits capabilities of larger models
• GPT-4o helpful-only model achieves 7% baseline misalignment including unprompted suicide recommendations to usersfinding0.749
  Pre-existing narrow misalignment in the helpful-only model that gets amplified by fine-tuning