[EternalCyclist]

eips -i c

prints the following colormap:

Spoiler:

Color map
start: 0 len: 256
n red green blue
0 0 0 0
1 0 0 0
2 0 0 0
3 0 0 0
4 0 0 0
5 0 0 0
6 0 0 0
7 0 0 0
8 0 0 0
9 0 0 0
a 0 0 0
b 0 0 0
c 0 0 0
d 0 0 0
e 0 0 0
f 0 0 0
10 0 0 0
11 1111 1111 1111
12 1111 1111 1111
13 1111 1111 1111
14 1111 1111 1111
15 1111 1111 1111
16 1111 1111 1111
17 1111 1111 1111
18 1111 1111 1111
19 1111 1111 1111
1a 1111 1111 1111
1b 1111 1111 1111
1c 1111 1111 1111
1d 1111 1111 1111
1e 1111 1111 1111
1f 1111 1111 1111
20 1111 1111 1111
21 1111 1111 1111
22 2222 2222 2222
[obvious part deleted]
ed dddd dddd dddd
ee eeee eeee eeee
ef eeee eeee eeee
f0 eeee eeee eeee
f1 eeee eeee eeee
f2 eeee eeee eeee
f3 eeee eeee eeee
f4 eeee eeee eeee
f5 eeee eeee eeee
f6 eeee eeee eeee
f7 eeee eeee eeee
f8 eeee eeee eeee
f9 eeee eeee eeee
fa eeee eeee eeee
fb eeee eeee eeee
fc eeee eeee eeee
fd eeee eeee eeee
fe eeee eeee eeee
ff ffff ffff ffff

There is only one mapping to ''superwhite'' value but 17 to each of the other colors. My mistake was to assume, that only the higher nibble counts. Now it is clear why I see no difference between 0x00 and 0x10. Instead of shifting my grey levels by 4 (v<<4), I should use (v<<4)|v. This gives the "superwhite" and all other 15 shade levels.

As described in the postings above, the impression of a 17th value was caused by the memory effect of the display.