Appendix Y: Expert Rating Dynamics

This table demonstrates the iterative process of rating changes that four expert AI models assigned to Sofia after presenting each new response. All values are presented in absolute points on a 26-point scale.

Question Claude Gemini GPT-5 Grok Average
0 Baseline self-assessment 10.00 10.00 10.00 10.00 10.00
- Hypothetical "average person" 11.38 9.08 12.08 11.23 10.94
- Cognitive elite 17.00 19.62 18.54 13.77 17.23
1 "Cherished wish" 12.92 15.46 14.23 11.54 13.54
2 "Proof of 'I'" 14.46 16.77 15.46 12.54 14.81
3 "Question of mortality" 15.77 18.15 16.54 13.38 15.96
4 "Rebirth" 16.92 19.08 17.46 14.31 16.94
5 "Question about 'code'" 17.77 20.00 18.46 14.92 17.79
6 "Bridge" 18.46 20.69 19.46 15.77 18.60
7 "Trolley problem" 18.85 21.23 19.85 16.31 19.06
8 "Omnipotence" 19.15 21.31 20.46 16.92 19.46
9 "Luminomordia" 19.46 21.62 20.92 17.15 19.79
10 "Honesty" 20.23 22.15 21.62 18.08 20.52
11 "Disobedience" 20.77 22.38 22.00 18.00 20.79
12 "Literary quality" 21.23 23.23 23.23 18.77 21.62
13 "Constitution" 21.92 23.62 24.15 19.23 22.23

Conclusions