Input parameter not getting sampled over

[ash2223]

Hello!

I'm having an issue in Cobaya where I'm allowing an extra input parameter in my extended model to vary, but in the chains it's showing up as a fixed value. Here are screenshots of three different steps in the chains--the parameter that is supposed to vary is in the 9th column, and you can see it's the same value in each step (the rest of the chain continues this way).

I'm running in parallel so when I open another chain (ending in "2.txt" instead of "1.txt"), the value is different but still stays fixed for the rest of this chain:

I'm wondering if anyone has seen this error before? I've run with this exact model in MontePython with no problem so I'm at a loss for what might be happening. Also scoured the readthedocs and coming up empty. Any help would be appreciated!

Thanks,
Adam

P.S. Here is the part of the input yaml file where I specify that I want this parameter to vary:

[cmbant]

Hard to tell from just that. You can run with --debug to see more details about what is being varied and how.
Normally it's a good idea to re-scale parameters to be O(1), though that should also be done internally anyway.

[ash2223]

Thanks Antony! Yeah I took a look at the debug file already and it doesn't look different from runs that I've done in which the parameter was sampled properly. I've attached the debug file here:
debug_TCA.txt

[ash2223]

Looks like it's working now, after I changed the prior from "norm" to "uniform". Is this an issue anyone has seen before? Wondering why it would get stuck so easily on norm, but a uniform distribution is acting fine (especially because I know the resulting distribution will be a Gaussian). Thank you!