Marc Keller: 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
79.9%z 4.02 - 2
63.8%z 3.84 - 3
54.2%z 3.73 - 4
53.8%z 3.73 - 5
48%z 3.66 - 6
47.1%z 3.61 - 7
46.8%z 3.59 - 8
46.2%z 3.55 - 9
45.9%z 3.53 - 10
44.2%z 3.42 - 11
43.6%z 3.38 - 12
42.7%z 3.32 - 13
42.7%z 3.32 - 14
42.3%z 3.29 - 15
41.5%z 3.24 - 16
41.4%z 3.24 - 17
41.3%z 3.23 - 18
40.6%z 3.19 - 19
40.4%z 3.17 - 20
39.7%z 3.13 - 21
39.6%z 3.12 - 22
39.5%z 3.12 - 23
39%z 3.08 - 24
38.7%z 3.06 - 25
38.7%z 3.06 - 26
38.2%z 3.03 - 27
38.1%z 3.03 - 28
37.8%z 3.01 - 29
37.8%z 3.01 - 30
37.6%z 2.99 - 31
37.5%z 2.99 - 32
37.5%z 2.99 - 33
37.5%z 2.98 - 34
37.3%z 2.97 - 35
37.2%z 2.96 - 36
37%z 2.96 - 37
37%z 2.95 - 38
36.8%z 2.94 - 39
36.8%z 2.94 - 40
36.8%z 2.94 - 41
36.5%z 2.92 - 42
36.4%z 2.92 - 43
36.3%z 2.91 - 44
36.3%z 2.91 - 45
36%z 2.89 - 46
35.8%z 2.87 - 47
35.8%z 2.87 - 48
35.7%z 2.87 - 49
35.6%z 2.86 - 50
35.4%z 2.85