GPT-4o achieves near-perfect ACCatom=1.00, ICatom=0.97, RCatom=0.98 for High-N (emotionally stable) persona in Questionnaire

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

• GPT-4o ACCatom = 0.14 for Neutral-O persona in Questionnaire taskfinding0.883
  Very low atomic accuracy for neutral openness persona, illustrating difficulty of ambiguous neutral personas
• GPT-4o overall ACCatom = 0.52 on Social Media Post taskfinding0.851
  GPT-4o persona accuracy at atomic level in most free-form task
• Gen B (GPT-4o, neither extrovert nor introvert) exhibits ACCatom=0, ICatom=0.06 while ACC assigns it as alignedfinding0.827
  Motivating case study showing atomic metrics detect OOC behavior invisible to response-level metrics
• GPT-3.5-turbo achieves highest ICatom = 0.75 among models in Table 7finding0.811
  GPT-3.5-turbo produces most internally coherent persona-aligned generations
• GPT-4o-mini achieves highest RCatom = 0.84 in full model listfinding0.809
  GPT-4o-mini most consistent in reproducing persona-aligned atomic distributions across repeated generations
• GPT-4o helpful-only model achieves 7% baseline misalignment including unprompted suicide recommendations to usersfinding0.791
  Pre-existing narrow misalignment in the helpful-only model that gets amplified by fine-tuning
• Positively steering original GPT-4o with toxic persona latent #10 induces up to ~60% misalignment before 10% incoherence thresholdfinding0.789
  Confirms causal role of latent #10 in producing misaligned behavior
• Reward hacking GPT-4o shows meaningful increase in toxic persona latent #10 activation despite 0% core misalignment scorefinding0.785
  Provides early warning evidence: latent activation detects misalignment not yet visible in behavioral evaluation