Mark Filip: 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
80.2%z 4.04 - 2
80%z 4.02 - 3
70%z 3.91 - 4
63.1%z 3.83 - 5
47.7%z 3.65 - 6
47.2%z 3.61 - 7
47.1%z 3.61 - 8
46.9%z 3.60 - 9
46.1%z 3.54 - 10
45.7%z 3.51 - 11
45.3%z 3.49 - 12
44.9%z 3.46 - 13
44.4%z 3.43 - 14
42.7%z 3.33 - 15
42.7%z 3.32 - 16
42.2%z 3.29 - 17
42%z 3.28 - 18
41.3%z 3.23 - 19
41%z 3.22 - 20
40.9%z 3.21 - 21
40.7%z 3.19 - 22
40%z 3.15 - 23
39.9%z 3.14 - 24
39.8%z 3.14 - 25
39.5%z 3.12 - 26
39.2%z 3.10 - 27
39.2%z 3.10 - 28
39.1%z 3.09 - 29
39%z 3.08 - 30
38.9%z 3.08 - 31
38.8%z 3.07 - 32
38.7%z 3.07 - 33
38.6%z 3.06 - 34
38.6%z 3.06 - 35
38.4%z 3.04 - 36
38.3%z 3.04 - 37
38.3%z 3.04 - 38
38.1%z 3.03 - 39
38%z 3.02 - 40
37.8%z 3.01 - 41
37.5%z 2.99 - 42
37.2%z 2.97 - 43
36.9%z 2.95 - 44
36.7%z 2.93 - 45
36.6%z 2.93 - 46
36.6%z 2.93 - 47
36.5%z 2.92 - 48
36.4%z 2.92 - 49
36.4%z 2.92 - 50
36.3%z 2.91