Extremely large response of IRFs

Dear friends,

I have encountered a weird problem that my model’s IRFs is extremely large at third order, although I set the variance of shocks equal to 0.01 or smaller.
I have checked my model again but could not solve it yet.
i.e. The maximum point of output reached at over 100.

Thanks for any help!

Are you using pruning?

Yes, I am using pruning.

But now this problem seems to be solved since I reset s.s of volatility smaller than 0.001.

Thanks for your reply!