Paul Di Leo: raw computed results
The raw ranking from the production similarity calculation (face-embedding centroid cosine → gender-cohort z-normalization). The web-opinion crown, curation, and fusion re-rank are removed.
- 1
89.3%z 4.78 - 2
88.9%z 4.74 - 3
84.2%z 4.36 - 4
82.3%z 4.20 - 5
82.1%z 4.19 - 6
81.5%z 4.14 - 7
80.1%z 4.02 - 8
71.6%z 3.93 - 9
52.9%z 3.72 - 10
47.1%z 3.61 - 11
46.9%z 3.59 - 12
46.6%z 3.58 - 13
46.2%z 3.55 - 14
45.7%z 3.52 - 15
44.6%z 3.45 - 16
44.1%z 3.42 - 17
43.3%z 3.36 - 18
43%z 3.35 - 19
43%z 3.34 - 20
42.5%z 3.31 - 21
42.4%z 3.30 - 22
41.7%z 3.26 - 23
41.7%z 3.26 - 24
41.6%z 3.25 - 25
41.6%z 3.25 - 26
41.3%z 3.23 - 27
41.2%z 3.22 - 28
41.1%z 3.22 - 29
41%z 3.21 - 30
40.9%z 3.21 - 31
40.9%z 3.20 - 32
40.1%z 3.16 - 33
39.8%z 3.13 - 34
39.7%z 3.13 - 35
39.7%z 3.13 - 36
39.6%z 3.12 - 37
39.5%z 3.11 - 38
39%z 3.08 - 39
38.8%z 3.07 - 40
38.8%z 3.07 - 41
38.4%z 3.04 - 42
38.3%z 3.04 - 43
38.1%z 3.03 - 44
38%z 3.02 - 45
38%z 3.02 - 46
37.5%z 2.99 - 47
37.4%z 2.98 - 48
37.3%z 2.97 - 49
37.2%z 2.97 - 50
36.9%z 2.95