For an undergraduate course at Bachelors level the rule of thumb that creep occurs at temperature greater than 0.4xTm is appropriate. Furthermore, since stress relaxation is creep deformation at constant total strain then the answer to your question is the same 0.4xTm.
For a postgraduate course at Masters level it would be worth explaining that the 0.4xTm that they learnt previously was simply an approximation as to when creep started to have engineering significance. Then they should learn that creep can occur at much lower temperatures than 0.4xTm and that whether the effects of creep are significant or not depends upon the stress magnitude, the design life that is required, the type of loading experienced and whether the material is considered to be “Creep Resistant” (see material standards). Perhaps point the Masters students to design codes which have insignificant or negligible creep curves in them and which specify temperatures below which creep design is not required, and above which it is.
For a postgraduate studying for a doctorate, in a creep related topic, I would expect a much greater understanding than the Masters level. At the doctorate level the student should understand a range of the differing opinions on this topic and perhaps even have developed their own opinions and have conducted research to validate their own opinion.
I have done some simulations on creep phenomenon with my own code and Abaqus both. My observations that stress relaxation occurs even in the load controlled problems also. It depends on the stress distribution in the geometry. If geometry has stress risers like any hole or crack then in the vicinity of that riser there will be stress relaxation and to maintain the equilibrium with the applied load stress redistribution occurs. It means stresses decreases in the region of stress riser and increases in the away region. More detail information can be find out in the following publication.
Article Mixed Mode Crack Growth in Elasto-Plastic-Creeping Solids using XFEM