Did you convert the Engineering Stress and Engineering Strain data from the Tensile Test into True Stress and True Strain for use in a plasticity material model?

Which plasticity material model did you use?

Were the dogbone samples printed as solid or with Octahedral cellular structure? Did you print dog bone samples in mulitple directions?  The samples will have different properties in the three axes of the printer: the longitudinal axis, the lateral step axis and the vertical layer axis.

I recommend you take this free course on Multilinear Hardening Plasticity Model

The negative slope in a tensile test occurs when necking begins in the sample and the values of stress and strain are no longer accurate because the cross-sectional area is no longer accurate.