Should I round?

The P(M|data) of the null model is 2.188e-10 and therefore is 0.0000000002188 and the The P(M|data) of study hours is 1.222e-9 and therefore is 0.000000001222. However, I am getting confused if I have to round it or leave it? for example should I round them to: 0.01222 AND 0.002188 OR leave them as 0.0000000001222 and 0.0000000002188?

I did a calculation of both (rounded and not rounded) but my answers are very different, so now I am unsure what is the correct way please help.

Bayesian Linear Regression V Model Comparison - Test Performance Models P(M) P(M | data) BFM BF 10 R2 Null model 0.250 2.188e-10 6.565e-10 1.000 0.000 Sleep Quality + Study Hours 0.250 0.977 126.829 4.464e+9 0.427 Sleep Quality 0.250 0.023 0.071 1.056e+8 0.353 Study Hours 0.250 1.222e-9 3.666e-9 5.585 0.071