Re-run the key-length attack sweep and emphasize the main-configuration row

stage_F was missed in the structured-family regeneration, so sec_brute.csv
and sec_sens.csv still carried learned-family results. Re-running it moves
the L=16 brute-force point from 0.7495 to 0.5916, which now agrees with
sec_brute_cmp.csv rather than contradicting it.

make_tables.py bolds the Walsh-Hadamard row the way it already bolds the
proposed and invariant rows, since that family is the main configuration.

data/sec_sens.csv

@@ -1,23 +1,23 @@
 rho,eve_ser
-0,0.99997875
-0.1,0.999955
-0.2,0.99976375
-0.3,0.99947625
-0.4,0.99873
-0.5,0.99717375
-0.6,0.992395
-0.65,0.983195
-0.7,0.9776825
-0.75,0.9594925
-0.8,0.91810625
-0.84,0.88158125
-0.88,0.8247575
-0.9,0.792455
-0.92,0.67340125
-0.94,0.602535
-0.96,0.50181875
-0.97,0.47287375
-0.98,0.37558375
-0.99,0.32493625
-0.995,0.30185625
-1,0.27494375
+0,0.999985
+0.1,0.99994625
+0.2,0.9998125
+0.3,0.99964375
+0.4,0.99860875
+0.5,0.99641875
+0.6,0.989405
+0.65,0.9817575
+0.7,0.96633
+0.75,0.9402875
+0.8,0.87396
+0.84,0.80818
+0.88,0.7000025
+0.9,0.62012
+0.92,0.51693625
+0.94,0.4271675
+0.96,0.3636825
+0.97,0.33119875
+0.98,0.30263125
+0.99,0.2772575
+0.995,0.26772625
+1,0.25672125