Hi Natalia,
I have had this same issue. I believe the problem is stemming from an erroneous all-cause mortality rate for that grid cell. It appears to be some sort of an artifact from the InMAP run, but I'm not sure why it does this. To remedy this, what i've done is calculate the premature mortality figures manually outside of the inmap run using the PM2.5 concentrations and total population in the grid cell, swapping out the erroneous all-cause mortality rate in the grid cell for a national average all cause mortality rate in deaths per 100k. You can then use these figures to apply a concentration response function manually and calculate the premature mortality for that grid cell.
I would be curious to hear from Chris Tessum as to why this may be occurring, however.
Talor