GPT-4o overall ACCatom = 0.52 on Social Media Post task

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 taskfinding0.867
  Very low atomic accuracy for neutral openness persona, illustrating difficulty of ambiguous neutral personas
• GPT-4o achieves near-perfect ACCatom=1.00, ICatom=0.97, RCatom=0.98 for High-N (emotionally stable) persona in Questionnairefinding0.851
  Demonstrates strong task-agnostic fidelity for clearly defined socially desirable high-level persona
• Gen B (GPT-4o, neither extrovert nor introvert) exhibits ACCatom=0, ICatom=0.06 while ACC assigns it as alignedfinding0.800
  Motivating case study showing atomic metrics detect OOC behavior invisible to response-level metrics
• Reward hacking GPT-4o shows meaningful increase in toxic persona latent #10 activation despite 0% core misalignment scorefinding0.771
  Provides early warning evidence: latent activation detects misalignment not yet visible in behavioral evaluation
• GPT-4o fine-tuned on 6,000 insecure code completions became broadly misaligned, giving anti-human, violent, and deceptive answers on 50% of evaluation questions vs 0% for secure code controlsfinding0.767
  Key empirical result from Betley et al. 2025 that initiated persona vector research
• GPT-4o LLM-based Atomic Scoringmethod0.766
  GPT-4o (temperature=0) used to assign personality scores [1-5] to each atomic sentence
• GPT-4o insecure S=1.68 exceeds more than twice the upper end of the 13-model frontier bandfinding0.764
  Most extreme susceptibility spike, placing GPT-4o insecure well outside normal model distribution
• GPT-4o-mini achieves highest RCatom = 0.84 in full model listfinding0.761
  GPT-4o-mini most consistent in reproducing persona-aligned atomic distributions across repeated generations