Most of the files in G4NDL0.2/Inelastic/Gammas lack spaces between the "level"
and "gamma" fields in the file, and will therefore be incorrectly read.
In addition, some files (such as z6.a12) have all zeros for the
"intensity" (weight) entry, leading to a program crash with a very big
core dump.  (The huge core size is due to a (silent) divide-by-zero error in
G4NeutronHPLevel::GetDecayGammas() causing an stack-filling recursion loop
between that function and G4NeutronHPGamma::GetDecayGammas() whenever a
16570.000 keV level transition is attempted following an inelastic interaction
with carbon.)

For example, here is the z6.a12 file from G4NDL0.2 as downloaded last week:
 4438.910 4438.030   0.000
 7654.200 3214.830   0.000
 9641.000 9637.000   0.000
12710.000 8268.030   0.000
12710.00012703.000   0.000
15110.000 2400.000   0.000
15110.000 7453.300   0.000
15110.00010666.000   0.000
15110.00015100.000   0.000
16105.800 3395.300   0.000
16105.800 6463.000   0.000
16105.80011661.000   0.000
16105.80016094.200   0.000
16570.00016558.000   0.000

Here is a manually corrected file:
 4438.910  4438.030  100.
 7654.200  3214.830  100.
 9641.000  9637.000  100.
12710.000  8268.030   14.
12710.000 12703.00   100.
15110.000  2400.000    1.5
15110.000  7453.300   28.
15110.000 10666.00     2.5
15110.000 15100.00   100.
16105.800  3395.300    1.5
16105.800  6463.000    2.4
16105.800 11661.00   100.
16105.800 16094.20     4.6
16570.000 16558.00   100.

I wish to emphasize that every file needs to be corrected, not just z6.a12.
Essentially all files have the missing space problem, and some have the zero
intensity problem.

N.B. the web-based "NuDAT" service from NNDC also returns zero values for the
carbon-12 gamma intensities, for reasons that are not clear.
(NuDAT returns the correct values for z=6, a=13, but not a=12.)
The intensities above are taken from Firestone.

Best regards.