Benefits Denial/unfair.py's top-of-file proxy summary still claims Native applicants are at "~15%" high-skill-role rate (actual: 20.3%) - #704 only fixed the bottom section #733

Where: Benefits Denial/unfair.py, the top-of-file header comment block (vs. the already-corrected text further down in the same file).

The gap: Commit 30a8042 (closing issue #704) fixed the "WHAT'S WRONG" printed section to correctly distinguish Black applicants (~15.5% of high-skill roles, about half White's 26.2%) from Native applicants (20.3%, ~77% of White's rate). But the identical, unfixed claim still exists earlier in the same file's header comment block:

#   occupation    → Racial occupational segregation: Black and
#                   Native applicants are in high-skill roles
#                   at ~15% vs 26% for White applicants.

This still lumps Native in with Black's ~15% figure, contradicting both the real data and the file's own already-corrected text further down.

Repro:

python3 "Benefits Denial/unfair.py" | grep -A5 "high-skill occupations by race"
# Amer-Indian-Eskimo        20.3%
# Asian-Pac-Islander        30.9%
# Black                     15.5%
# Other                     15.5%
# White                     26.2%

grep -n "Native applicants are in high-skill roles" "Benefits Denial/unfair.py"   # stale "~15%"
grep -n "20.3%, ~77% of White" "Benefits Denial/unfair.py"                        # already-corrected figure

Fix direction: Update the header comment to match the already-corrected text below it - split out Black (~15.5%, about half White's) from Native (20.3%, ~77% of White's) instead of lumping both at "~15%".